5 Changes between 0.9.8x and 0.9.8y [xx XXX xxxx]
9 Changes between 0.9.8w and 0.9.8x [10 May 2012]
11 *) Sanity check record length before skipping explicit IV in DTLS
14 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
15 fuzzing as a service testing platform.
19 *) Initialise tkeylen properly when encrypting CMS messages.
20 Thanks to Solar Designer of Openwall for reporting this issue.
23 Changes between 0.9.8v and 0.9.8w [23 Apr 2012]
25 *) The fix for CVE-2012-2110 did not take into account that the
26 'len' argument to BUF_MEM_grow and BUF_MEM_grow_clean is an
27 int in OpenSSL 0.9.8, making it still vulnerable. Fix by
28 rejecting negative len parameter. (CVE-2012-2131)
29 [Tomas Hoger <thoger@redhat.com>]
31 Changes between 0.9.8u and 0.9.8v [19 Apr 2012]
33 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
34 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
35 in CRYPTO_realloc_clean.
37 Thanks to Tavis Ormandy, Google Security Team, for discovering this
38 issue and to Adam Langley <agl@chromium.org> for fixing it.
40 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
42 Changes between 0.9.8t and 0.9.8u [12 Mar 2012]
44 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
45 in CMS and PKCS7 code. When RSA decryption fails use a random key for
46 content decryption and always return the same error. Note: this attack
47 needs on average 2^20 messages so it only affects automated senders. The
48 old behaviour can be reenabled in the CMS code by setting the
49 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
50 an MMA defence is not necessary.
51 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
52 this issue. (CVE-2012-0884)
55 *) Fix CVE-2011-4619: make sure we really are receiving a
56 client hello before rejecting multiple SGC restarts. Thanks to
57 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
60 Changes between 0.9.8s and 0.9.8t [18 Jan 2012]
62 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
63 Thanks to Antonio Martin, Enterprise Secure Access Research and
64 Development, Cisco Systems, Inc. for discovering this bug and
65 preparing a fix. (CVE-2012-0050)
68 Changes between 0.9.8r and 0.9.8s [4 Jan 2012]
70 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
71 of the Vaudenay padding oracle attack on CBC mode encryption
72 which enables an efficient plaintext recovery attack against
73 the OpenSSL implementation of DTLS. Their attack exploits timing
74 differences arising during decryption processing. A research
75 paper describing this attack can be found at:
76 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
77 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
78 Security Group at Royal Holloway, University of London
79 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
80 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
81 for preparing the fix. (CVE-2011-4108)
82 [Robin Seggelmann, Michael Tuexen]
84 *) Stop policy check failure freeing same buffer twice. (CVE-2011-4109)
85 [Ben Laurie, Kasper <ekasper@google.com>]
87 *) Clear bytes used for block padding of SSL 3.0 records.
89 [Adam Langley (Google)]
91 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
92 Kadianakis <desnacked@gmail.com> for discovering this issue and
93 Adam Langley for preparing the fix. (CVE-2011-4619)
94 [Adam Langley (Google)]
96 *) Prevent malformed RFC3779 data triggering an assertion failure.
97 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
98 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
99 [Rob Austein <sra@hactrn.net>]
101 *) Fix ssl_ciph.c set-up race.
102 [Adam Langley (Google)]
104 *) Fix spurious failures in ecdsatest.c.
105 [Emilia Käsper (Google)]
107 *) Fix the BIO_f_buffer() implementation (which was mixing different
108 interpretations of the '..._len' fields).
109 [Adam Langley (Google)]
111 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
112 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
113 threads won't reuse the same blinding coefficients.
115 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
116 lock to call BN_BLINDING_invert_ex, and avoids one use of
117 BN_BLINDING_update for each BN_BLINDING structure (previously,
118 the last update always remained unused).
119 [Emilia Käsper (Google)]
121 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
122 for multi-threaded use of ECDH.
123 [Adam Langley (Google)]
125 *) Fix x509_name_ex_d2i memory leak on bad inputs.
128 *) Add protection against ECDSA timing attacks as mentioned in the paper
129 by Billy Bob Brumley and Nicola Tuveri, see:
131 http://eprint.iacr.org/2011/232.pdf
133 [Billy Bob Brumley and Nicola Tuveri]
135 Changes between 0.9.8q and 0.9.8r [8 Feb 2011]
137 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
138 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
140 *) Fix bug in string printing code: if *any* escaping is enabled we must
141 escape the escape character (backslash) or the resulting string is
145 Changes between 0.9.8p and 0.9.8q [2 Dec 2010]
147 *) Disable code workaround for ancient and obsolete Netscape browsers
148 and servers: an attacker can use it in a ciphersuite downgrade attack.
149 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
152 *) Fixed J-PAKE implementation error, originally discovered by
153 Sebastien Martini, further info and confirmation from Stefan
154 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
157 Changes between 0.9.8o and 0.9.8p [16 Nov 2010]
159 *) Fix extension code to avoid race conditions which can result in a buffer
160 overrun vulnerability: resumed sessions must not be modified as they can
161 be shared by multiple threads. CVE-2010-3864
164 *) Fix for double free bug in ssl/s3_clnt.c CVE-2010-2939
167 *) Don't reencode certificate when calculating signature: cache and use
168 the original encoding instead. This makes signature verification of
169 some broken encodings work correctly.
172 *) ec2_GF2m_simple_mul bugfix: compute correct result if the output EC_POINT
173 is also one of the inputs.
174 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
176 *) Don't repeatedly append PBE algorithms to table if they already exist.
177 Sort table on each new add. This effectively makes the table read only
178 after all algorithms are added and subsequent calls to PKCS12_pbe_add
182 Changes between 0.9.8n and 0.9.8o [01 Jun 2010]
184 [NB: OpenSSL 0.9.8o and later 0.9.8 patch levels were released after
187 *) Correct a typo in the CMS ASN1 module which can result in invalid memory
188 access or freeing data twice (CVE-2010-0742)
189 [Steve Henson, Ronald Moesbergen <intercommit@gmail.com>]
191 *) Add SHA2 algorithms to SSL_library_init(). SHA2 is becoming far more
192 common in certificates and some applications which only call
193 SSL_library_init and not OpenSSL_add_all_algorithms() will fail.
197 Reduce copying into .apps and .test in makevms.com
198 Don't try to use blank CA certificate in CA.com
199 Allow use of C files from original directories in maketests.com
200 [Steven M. Schweda" <sms@antinode.info>]
202 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
204 *) When rejecting SSL/TLS records due to an incorrect version number, never
205 update s->server with a new major version number. As of
206 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
207 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
208 the previous behavior could result in a read attempt at NULL when
209 receiving specific incorrect SSL/TLS records once record payload
210 protection is active. (CVE-2010-0740)
211 [Bodo Moeller, Adam Langley <agl@chromium.org>]
213 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
214 could be crashed if the relevant tables were not present (e.g. chrooted).
215 [Tomas Hoger <thoger@redhat.com>]
217 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
219 *) Always check bn_wexpend() return values for failure. (CVE-2009-3245)
220 [Martin Olsson, Neel Mehta]
222 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
223 accommodate for stack sorting, always a write lock!).
226 *) On some versions of WIN32 Heap32Next is very slow. This can cause
227 excessive delays in the RAND_poll(): over a minute. As a workaround
228 include a time check in the inner Heap32Next loop too.
231 *) The code that handled flushing of data in SSL/TLS originally used the
232 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
233 the problem outlined in PR#1949. The fix suggested there however can
234 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
235 of Apache). So instead simplify the code to flush unconditionally.
236 This should be fine since flushing with no data to flush is a no op.
239 *) Handle TLS versions 2.0 and later properly and correctly use the
240 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
241 off ancient servers have a habit of sticking around for a while...
244 *) Modify compression code so it frees up structures without using the
245 ex_data callbacks. This works around a problem where some applications
246 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
247 restarting) then use compression (e.g. SSL with compression) later.
248 This results in significant per-connection memory leaks and
249 has caused some security issues including CVE-2008-1678 and
253 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
254 change when encrypting or decrypting.
257 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
258 connect and renegotiate with servers which do not support RI.
259 Until RI is more widely deployed this option is enabled by default.
262 *) Add "missing" ssl ctrls to clear options and mode.
265 *) If client attempts to renegotiate and doesn't support RI respond with
266 a no_renegotiation alert as required by RFC5746. Some renegotiating
267 TLS clients will continue a connection gracefully when they receive
268 the alert. Unfortunately OpenSSL mishandled this alert and would hang
269 waiting for a server hello which it will never receive. Now we treat a
270 received no_renegotiation alert as a fatal error. This is because
271 applications requesting a renegotiation might well expect it to succeed
272 and would have no code in place to handle the server denying it so the
273 only safe thing to do is to terminate the connection.
276 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
277 peer supports secure renegotiation and 0 otherwise. Print out peer
278 renegotiation support in s_client/s_server.
281 *) Replace the highly broken and deprecated SPKAC certification method with
282 the updated NID creation version. This should correctly handle UTF8.
285 *) Implement RFC5746. Re-enable renegotiation but require the extension
286 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
287 turns out to be a bad idea. It has been replaced by
288 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
289 SSL_CTX_set_options(). This is really not recommended unless you
290 know what you are doing.
291 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
293 *) Fixes to stateless session resumption handling. Use initial_ctx when
294 issuing and attempting to decrypt tickets in case it has changed during
295 servername handling. Use a non-zero length session ID when attempting
296 stateless session resumption: this makes it possible to determine if
297 a resumption has occurred immediately after receiving server hello
298 (several places in OpenSSL subtly assume this) instead of later in
302 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
303 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
304 fixes for a few places where the return code is not checked
306 [Julia Lawall <julia@diku.dk>]
308 *) Add --strict-warnings option to Configure script to include devteam
309 warnings in other configurations.
312 *) Add support for --libdir option and LIBDIR variable in makefiles. This
313 makes it possible to install openssl libraries in locations which
314 have names other than "lib", for example "/usr/lib64" which some
316 [Steve Henson, based on patch from Jeremy Utley]
318 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
319 X690 8.9.12 and can produce some misleading textual output of OIDs.
320 [Steve Henson, reported by Dan Kaminsky]
322 *) Delete MD2 from algorithm tables. This follows the recommendation in
323 several standards that it is not used in new applications due to
324 several cryptographic weaknesses. For binary compatibility reasons
325 the MD2 API is still compiled in by default.
328 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
332 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
333 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
335 [Guenter <lists@gknw.net>]
337 *) Fix the server certificate chain building code to use X509_verify_cert(),
338 it used to have an ad-hoc builder which was unable to cope with anything
339 other than a simple chain.
340 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
342 *) Don't check self signed certificate signatures in X509_verify_cert()
343 by default (a flag can override this): it just wastes time without
344 adding any security. As a useful side effect self signed root CAs
345 with non-FIPS digests are now usable in FIPS mode.
348 *) In dtls1_process_out_of_seq_message() the check if the current message
349 is already buffered was missing. For every new message was memory
350 allocated, allowing an attacker to perform an denial of service attack
351 with sending out of seq handshake messages until there is no memory
352 left. Additionally every future messege was buffered, even if the
353 sequence number made no sense and would be part of another handshake.
354 So only messages with sequence numbers less than 10 in advance will be
355 buffered. (CVE-2009-1378)
356 [Robin Seggelmann, discovered by Daniel Mentz]
358 *) Records are buffered if they arrive with a future epoch to be
359 processed after finishing the corresponding handshake. There is
360 currently no limitation to this buffer allowing an attacker to perform
361 a DOS attack with sending records with future epochs until there is no
362 memory left. This patch adds the pqueue_size() function to detemine
363 the size of a buffer and limits the record buffer to 100 entries.
365 [Robin Seggelmann, discovered by Daniel Mentz]
367 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
368 parent structure is freed. (CVE-2009-1379)
371 *) Handle non-blocking I/O properly in SSL_shutdown() call.
372 [Darryl Miles <darryl-mailinglists@netbauds.net>]
375 [Ilya O. <vrghost@gmail.com>]
377 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
379 *) Disable renegotiation completely - this fixes a severe security
380 problem (CVE-2009-3555) at the cost of breaking all
381 renegotiation. Renegotiation can be re-enabled by setting
382 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
383 run-time. This is really not recommended unless you know what
387 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
389 *) Don't set val to NULL when freeing up structures, it is freed up by
390 underlying code. If sizeof(void *) > sizeof(long) this can result in
391 zeroing past the valid field. (CVE-2009-0789)
392 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
394 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
395 checked correctly. This would allow some invalid signed attributes to
396 appear to verify correctly. (CVE-2009-0591)
397 [Ivan Nestlerode <inestlerode@us.ibm.com>]
399 *) Reject UniversalString and BMPString types with invalid lengths. This
400 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
401 a legal length. (CVE-2009-0590)
404 *) Set S/MIME signing as the default purpose rather than setting it
405 unconditionally. This allows applications to override it at the store
409 *) Permit restricted recursion of ASN1 strings. This is needed in practice
410 to handle some structures.
413 *) Improve efficiency of mem_gets: don't search whole buffer each time
415 [Jeremy Shapiro <jnshapir@us.ibm.com>]
417 *) New -hex option for openssl rand.
420 *) Print out UTF8String and NumericString when parsing ASN1.
423 *) Support NumericString type for name components.
426 *) Allow CC in the environment to override the automatically chosen
427 compiler. Note that nothing is done to ensure flags work with the
431 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
433 *) Properly check EVP_VerifyFinal() and similar return values
435 [Ben Laurie, Bodo Moeller, Google Security Team]
437 *) Enable TLS extensions by default.
440 *) Allow the CHIL engine to be loaded, whether the application is
441 multithreaded or not. (This does not release the developer from the
442 obligation to set up the dynamic locking callbacks.)
443 [Sander Temme <sander@temme.net>]
445 *) Use correct exit code if there is an error in dgst command.
446 [Steve Henson; problem pointed out by Roland Dirlewanger]
448 *) Tweak Configure so that you need to say "experimental-jpake" to enable
449 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
452 *) Add experimental JPAKE support, including demo authentication in
453 s_client and s_server.
456 *) Set the comparison function in v3_addr_canonize().
457 [Rob Austein <sra@hactrn.net>]
459 *) Add support for XMPP STARTTLS in s_client.
460 [Philip Paeps <philip@freebsd.org>]
462 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
463 to ensure that even with this option, only ciphersuites in the
464 server's preference list will be accepted. (Note that the option
465 applies only when resuming a session, so the earlier behavior was
466 just about the algorithm choice for symmetric cryptography.)
469 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
471 *) Fix NULL pointer dereference if a DTLS server received
472 ChangeCipherSpec as first record (CVE-2009-1386).
475 *) Fix a state transitition in s3_srvr.c and d1_srvr.c
476 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
479 *) The fix in 0.9.8c that supposedly got rid of unsafe
480 double-checked locking was incomplete for RSA blinding,
481 addressing just one layer of what turns out to have been
482 doubly unsafe triple-checked locking.
484 So now fix this for real by retiring the MONT_HELPER macro
485 in crypto/rsa/rsa_eay.c.
487 [Bodo Moeller; problem pointed out by Marius Schilder]
489 *) Various precautionary measures:
491 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
493 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
494 (NB: This would require knowledge of the secret session ticket key
495 to exploit, in which case you'd be SOL either way.)
497 - Change bn_nist.c so that it will properly handle input BIGNUMs
498 outside the expected range.
500 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
503 [Neel Mehta, Bodo Moeller]
505 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
506 the load fails. Useful for distros.
507 [Ben Laurie and the FreeBSD team]
509 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
512 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
515 *) Expand ENGINE to support engine supplied SSL client certificate functions.
517 This work was sponsored by Logica.
520 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
521 keystores. Support for SSL/TLS client authentication too.
522 Not compiled unless enable-capieng specified to Configure.
524 This work was sponsored by Logica.
527 *) Fix bug in X509_ATTRIBUTE creation: dont set attribute using
528 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
529 attribute creation routines such as certifcate requests and PKCS#12
533 Changes between 0.9.8g and 0.9.8h [28 May 2008]
535 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
536 handshake which could lead to a cilent crash as found using the
537 Codenomicon TLS test suite (CVE-2008-1672)
538 [Steve Henson, Mark Cox]
540 *) Fix double free in TLS server name extensions which could lead to
541 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
544 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
546 Clear the error queue to ensure that error entries left from
547 older function calls do not interfere with the correct operation.
548 [Lutz Jaenicke, Erik de Castro Lopo]
550 *) Remove root CA certificates of commercial CAs:
552 The OpenSSL project does not recommend any specific CA and does not
553 have any policy with respect to including or excluding any CA.
554 Therefore it does not make any sense to ship an arbitrary selection
555 of root CA certificates with the OpenSSL software.
558 *) RSA OAEP patches to fix two separate invalid memory reads.
559 The first one involves inputs when 'lzero' is greater than
560 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
561 before the beginning of from). The second one involves inputs where
562 the 'db' section contains nothing but zeroes (there is a one-byte
563 invalid read after the end of 'db').
564 [Ivan Nestlerode <inestlerode@us.ibm.com>]
566 *) Partial backport from 0.9.9-dev:
568 Introduce bn_mul_mont (dedicated Montgomery multiplication
569 procedure) as a candidate for BIGNUM assembler implementation.
570 While 0.9.9-dev uses assembler for various architectures, only
571 x86_64 is available by default here in the 0.9.8 branch, and
572 32-bit x86 is available through a compile-time setting.
574 To try the 32-bit x86 assembler implementation, use Configure
575 option "enable-montasm" (which exists only for this backport).
577 As "enable-montasm" for 32-bit x86 disclaims code stability
578 anyway, in this constellation we activate additional code
579 backported from 0.9.9-dev for further performance improvements,
580 namely BN_from_montgomery_word. (To enable this otherwise,
581 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
583 [Andy Polyakov (backport partially by Bodo Moeller)]
585 *) Add TLS session ticket callback. This allows an application to set
586 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
587 values. This is useful for key rollover for example where several key
588 sets may exist with different names.
591 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
592 This was broken until now in 0.9.8 releases, such that the only way
593 a registered ENGINE could be used (assuming it initialises
594 successfully on the host) was to explicitly set it as the default
595 for the relevant algorithms. This is in contradiction with 0.9.7
596 behaviour and the documentation. With this fix, when an ENGINE is
597 registered into a given algorithm's table of implementations, the
598 'uptodate' flag is reset so that auto-discovery will be used next
599 time a new context for that algorithm attempts to select an
601 [Ian Lister (tweaked by Geoff Thorpe)]
603 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
604 implemention in the following ways:
606 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
609 Lack of BER streaming support means one pass streaming processing is
610 only supported if data is detached: setting the streaming flag is
611 ignored for embedded content.
613 CMS support is disabled by default and must be explicitly enabled
614 with the enable-cms configuration option.
617 *) Update the GMP engine glue to do direct copies between BIGNUM and
618 mpz_t when openssl and GMP use the same limb size. Otherwise the
619 existing "conversion via a text string export" trick is still used.
620 [Paul Sheer <paulsheer@gmail.com>]
622 *) Zlib compression BIO. This is a filter BIO which compressed and
623 uncompresses any data passed through it.
626 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
627 RFC3394 compatible AES key wrapping.
630 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
631 sets string data without copying. X509_ALGOR_set0() and
632 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
633 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
634 from an X509_ATTRIBUTE structure optionally checking it occurs only
635 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
639 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
640 to get the expected BN_FLG_CONSTTIME behavior.
641 [Bodo Moeller (Google)]
645 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
646 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
647 - added some more tests to do_tests.pl
648 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
649 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
650 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
651 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
652 - various changes to netware.pl to enable gcc-cross builds on Win32
654 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
655 - various changes to fix missing prototype warnings
656 - fixed x86nasm.pl to create correct asm files for NASM COFF output
657 - added AES, WHIRLPOOL and CPUID assembler code to build files
658 - added missing AES assembler make rules to mk1mf.pl
659 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
660 [Guenter Knauf <eflash@gmx.net>]
662 *) Implement certificate status request TLS extension defined in RFC3546.
663 A client can set the appropriate parameters and receive the encoded
664 OCSP response via a callback. A server can query the supplied parameters
665 and set the encoded OCSP response in the callback. Add simplified examples
666 to s_client and s_server.
669 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
672 + Binary incompatibility of ssl_ctx_st structure
673 + DTLS interoperation with non-compliant servers
674 + Don't call get_session_cb() without proposed session
675 + Fix ia64 assembler code
676 [Andy Polyakov, Steve Henson]
678 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
680 *) DTLS Handshake overhaul. There were longstanding issues with
681 OpenSSL DTLS implementation, which were making it impossible for
682 RFC 4347 compliant client to communicate with OpenSSL server.
683 Unfortunately just fixing these incompatibilities would "cut off"
684 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
685 server keeps tolerating non RFC compliant syntax. The opposite is
686 not true, 0.9.8f client can not communicate with earlier server.
687 This update even addresses CVE-2007-4995.
690 *) Changes to avoid need for function casts in OpenSSL: some compilers
691 (gcc 4.2 and later) reject their use.
692 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
695 *) Add RFC4507 support to OpenSSL. This includes the corrections in
696 RFC4507bis. The encrypted ticket format is an encrypted encoded
697 SSL_SESSION structure, that way new session features are automatically
700 If a client application caches session in an SSL_SESSION structure
701 support is transparent because tickets are now stored in the encoded
704 The SSL_CTX structure automatically generates keys for ticket
705 protection in servers so again support should be possible
706 with no application modification.
708 If a client or server wishes to disable RFC4507 support then the option
709 SSL_OP_NO_TICKET can be set.
711 Add a TLS extension debugging callback to allow the contents of any client
712 or server extensions to be examined.
714 This work was sponsored by Google.
717 *) Add initial support for TLS extensions, specifically for the server_name
718 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
719 have new members for a host name. The SSL data structure has an
720 additional member SSL_CTX *initial_ctx so that new sessions can be
721 stored in that context to allow for session resumption, even after the
722 SSL has been switched to a new SSL_CTX in reaction to a client's
723 server_name extension.
725 New functions (subject to change):
728 SSL_get_servername_type()
731 New CTRL codes and macros (subject to change):
733 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
734 - SSL_CTX_set_tlsext_servername_callback()
735 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
736 - SSL_CTX_set_tlsext_servername_arg()
737 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
739 openssl s_client has a new '-servername ...' option.
741 openssl s_server has new options '-servername_host ...', '-cert2 ...',
742 '-key2 ...', '-servername_fatal' (subject to change). This allows
743 testing the HostName extension for a specific single host name ('-cert'
744 and '-key' remain fallbacks for handshakes without HostName
745 negotiation). If the unrecogninzed_name alert has to be sent, this by
746 default is a warning; it becomes fatal with the '-servername_fatal'
749 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
751 *) Add AES and SSE2 assembly language support to VC++ build.
754 *) Mitigate attack on final subtraction in Montgomery reduction.
757 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
758 (which previously caused an internal error).
761 *) Squeeze another 10% out of IGE mode when in != out.
764 *) AES IGE mode speedup.
765 [Dean Gaudet (Google)]
767 *) Add the Korean symmetric 128-bit cipher SEED (see
768 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
769 add SEED ciphersuites from RFC 4162:
771 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
772 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
773 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
774 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
776 To minimize changes between patchlevels in the OpenSSL 0.9.8
777 series, SEED remains excluded from compilation unless OpenSSL
778 is configured with 'enable-seed'.
781 *) Mitigate branch prediction attacks, which can be practical if a
782 single processor is shared, allowing a spy process to extract
783 information. For detailed background information, see
784 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
785 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
786 and Necessary Software Countermeasures"). The core of the change
787 are new versions BN_div_no_branch() and
788 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
789 respectively, which are slower, but avoid the security-relevant
790 conditional branches. These are automatically called by BN_div()
791 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
792 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
793 remove a conditional branch.
795 BN_FLG_CONSTTIME is the new name for the previous
796 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
797 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
798 in the exponent causes BN_mod_exp_mont() to use the alternative
799 implementation in BN_mod_exp_mont_consttime().) The old name
800 remains as a deprecated alias.
802 Similary, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
803 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
804 constant-time implementations for more than just exponentiation.
805 Here too the old name is kept as a deprecated alias.
807 BN_BLINDING_new() will now use BN_dup() for the modulus so that
808 the BN_BLINDING structure gets an independent copy of the
809 modulus. This means that the previous "BIGNUM *m" argument to
810 BN_BLINDING_new() and to BN_BLINDING_create_param() now
811 essentially becomes "const BIGNUM *m", although we can't actually
812 change this in the header file before 0.9.9. It allows
813 RSA_setup_blinding() to use BN_with_flags() on the modulus to
814 enable BN_FLG_CONSTTIME.
816 [Matthew D Wood (Intel Corp)]
818 *) In the SSL/TLS server implementation, be strict about session ID
819 context matching (which matters if an application uses a single
820 external cache for different purposes). Previously,
821 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
822 set. This did ensure strict client verification, but meant that,
823 with applications using a single external cache for quite
824 different requirements, clients could circumvent ciphersuite
825 restrictions for a given session ID context by starting a session
826 in a different context.
829 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
830 a ciphersuite string such as "DEFAULT:RSA" cannot enable
831 authentication-only ciphersuites.
834 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
835 not complete and could lead to a possible single byte overflow
836 (CVE-2007-5135) [Ben Laurie]
838 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
840 *) Since AES128 and AES256 (and similarly Camellia128 and
841 Camellia256) share a single mask bit in the logic of
842 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
843 kludge to work properly if AES128 is available and AES256 isn't
844 (or if Camellia128 is available and Camellia256 isn't).
847 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
848 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
849 When a point or a seed is encoded in a BIT STRING, we need to
850 prevent the removal of trailing zero bits to get the proper DER
851 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
852 of a NamedBitList, for which trailing 0 bits need to be removed.)
855 *) Have SSL/TLS server implementation tolerate "mismatched" record
856 protocol version while receiving ClientHello even if the
857 ClientHello is fragmented. (The server can't insist on the
858 particular protocol version it has chosen before the ServerHello
859 message has informed the client about his choice.)
862 *) Add RFC 3779 support.
863 [Rob Austein for ARIN, Ben Laurie]
865 *) Load error codes if they are not already present instead of using a
866 static variable. This allows them to be cleanly unloaded and reloaded.
867 Improve header file function name parsing.
870 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
871 or CAPABILITY handshake as required by RFCs.
874 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
876 *) Introduce limits to prevent malicious keys being able to
877 cause a denial of service. (CVE-2006-2940)
878 [Steve Henson, Bodo Moeller]
880 *) Fix ASN.1 parsing of certain invalid structures that can result
881 in a denial of service. (CVE-2006-2937) [Steve Henson]
883 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
884 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
886 *) Fix SSL client code which could crash if connecting to a
887 malicious SSLv2 server. (CVE-2006-4343)
888 [Tavis Ormandy and Will Drewry, Google Security Team]
890 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
891 match only those. Before that, "AES256-SHA" would be interpreted
892 as a pattern and match "AES128-SHA" too (since AES128-SHA got
893 the same strength classification in 0.9.7h) as we currently only
894 have a single AES bit in the ciphersuite description bitmap.
895 That change, however, also applied to ciphersuite strings such as
896 "RC4-MD5" that intentionally matched multiple ciphersuites --
897 namely, SSL 2.0 ciphersuites in addition to the more common ones
898 from SSL 3.0/TLS 1.0.
900 So we change the selection algorithm again: Naming an explicit
901 ciphersuite selects this one ciphersuite, and any other similar
902 ciphersuite (same bitmap) from *other* protocol versions.
903 Thus, "RC4-MD5" again will properly select both the SSL 2.0
904 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
906 Since SSL 2.0 does not have any ciphersuites for which the
907 128/256 bit distinction would be relevant, this works for now.
908 The proper fix will be to use different bits for AES128 and
909 AES256, which would have avoided the problems from the beginning;
910 however, bits are scarce, so we can only do this in a new release
911 (not just a patchlevel) when we can change the SSL_CIPHER
912 definition to split the single 'unsigned long mask' bitmap into
913 multiple values to extend the available space.
917 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
919 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
920 (CVE-2006-4339) [Ben Laurie and Google Security Team]
922 *) Add AES IGE and biIGE modes.
925 *) Change the Unix randomness entropy gathering to use poll() when
926 possible instead of select(), since the latter has some
927 undesirable limitations.
928 [Darryl Miles via Richard Levitte and Bodo Moeller]
930 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
931 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
932 cannot be implicitly activated as part of, e.g., the "AES" alias.
933 However, please upgrade to OpenSSL 0.9.9[-dev] for
934 non-experimental use of the ECC ciphersuites to get TLS extension
935 support, which is required for curve and point format negotiation
936 to avoid potential handshake problems.
939 *) Disable rogue ciphersuites:
941 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
942 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
943 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
945 The latter two were purportedly from
946 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
949 Also deactivate the remaining ciphersuites from
950 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
951 unofficial, and the ID has long expired.
954 *) Fix RSA blinding Heisenbug (problems sometimes occured on
955 dual-core machines) and other potential thread-safety issues.
958 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
959 versions), which is now available for royalty-free use
960 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
961 Also, add Camellia TLS ciphersuites from RFC 4132.
963 To minimize changes between patchlevels in the OpenSSL 0.9.8
964 series, Camellia remains excluded from compilation unless OpenSSL
965 is configured with 'enable-camellia'.
968 *) Disable the padding bug check when compression is in use. The padding
969 bug check assumes the first packet is of even length, this is not
970 necessarily true if compresssion is enabled and can result in false
971 positives causing handshake failure. The actual bug test is ancient
972 code so it is hoped that implementations will either have fixed it by
973 now or any which still have the bug do not support compression.
976 Changes between 0.9.8a and 0.9.8b [04 May 2006]
978 *) When applying a cipher rule check to see if string match is an explicit
979 cipher suite and only match that one cipher suite if it is.
982 *) Link in manifests for VC++ if needed.
983 [Austin Ziegler <halostatue@gmail.com>]
985 *) Update support for ECC-based TLS ciphersuites according to
986 draft-ietf-tls-ecc-12.txt with proposed changes (but without
987 TLS extensions, which are supported starting with the 0.9.9
988 branch, not in the OpenSSL 0.9.8 branch).
991 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
992 opaque EVP_CIPHER_CTX handling.
995 *) Fixes and enhancements to zlib compression code. We now only use
996 "zlib1.dll" and use the default __cdecl calling convention on Win32
997 to conform with the standards mentioned here:
998 http://www.zlib.net/DLL_FAQ.txt
999 Static zlib linking now works on Windows and the new --with-zlib-include
1000 --with-zlib-lib options to Configure can be used to supply the location
1001 of the headers and library. Gracefully handle case where zlib library
1005 *) Several fixes and enhancements to the OID generation code. The old code
1006 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
1007 handle numbers larger than ULONG_MAX, truncated printing and had a
1008 non standard OBJ_obj2txt() behaviour.
1011 *) Add support for building of engines under engine/ as shared libraries
1012 under VC++ build system.
1015 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
1016 Hopefully, we will not see any false combination of paths any more.
1019 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
1021 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
1022 (part of SSL_OP_ALL). This option used to disable the
1023 countermeasure against man-in-the-middle protocol-version
1024 rollback in the SSL 2.0 server implementation, which is a bad
1025 idea. (CVE-2005-2969)
1027 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
1028 for Information Security, National Institute of Advanced Industrial
1029 Science and Technology [AIST], Japan)]
1031 *) Add two function to clear and return the verify parameter flags.
1034 *) Keep cipherlists sorted in the source instead of sorting them at
1035 runtime, thus removing the need for a lock.
1038 *) Avoid some small subgroup attacks in Diffie-Hellman.
1039 [Nick Mathewson and Ben Laurie]
1041 *) Add functions for well-known primes.
1044 *) Extended Windows CE support.
1045 [Satoshi Nakamura and Andy Polyakov]
1047 *) Initialize SSL_METHOD structures at compile time instead of during
1048 runtime, thus removing the need for a lock.
1051 *) Make PKCS7_decrypt() work even if no certificate is supplied by
1052 attempting to decrypt each encrypted key in turn. Add support to
1056 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
1058 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
1061 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
1064 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
1065 key into the same file any more.
1068 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
1071 *) Add -utf8 command line and config file option to 'ca'.
1072 [Stefan <stf@udoma.org]
1074 *) Removed the macro des_crypt(), as it seems to conflict with some
1075 libraries. Use DES_crypt().
1078 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
1079 involves renaming the source and generated shared-libs for
1080 both. The engines will accept the corrected or legacy ids
1081 ('ncipher' and '4758_cca' respectively) when binding. NB,
1082 this only applies when building 'shared'.
1083 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
1085 *) Add attribute functions to EVP_PKEY structure. Modify
1086 PKCS12_create() to recognize a CSP name attribute and
1087 use it. Make -CSP option work again in pkcs12 utility.
1090 *) Add new functionality to the bn blinding code:
1091 - automatic re-creation of the BN_BLINDING parameters after
1092 a fixed number of uses (currently 32)
1093 - add new function for parameter creation
1094 - introduce flags to control the update behaviour of the
1095 BN_BLINDING parameters
1096 - hide BN_BLINDING structure
1097 Add a second BN_BLINDING slot to the RSA structure to improve
1098 performance when a single RSA object is shared among several
1102 *) Add support for DTLS.
1103 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
1105 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
1106 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
1109 *) Remove buggy and incompletet DH cert support from
1110 ssl/ssl_rsa.c and ssl/s3_both.c
1113 *) Use SHA-1 instead of MD5 as the default digest algorithm for
1114 the apps/openssl applications.
1117 *) Compile clean with "-Wall -Wmissing-prototypes
1118 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
1119 DEBUG_SAFESTACK must also be set.
1122 *) Change ./Configure so that certain algorithms can be disabled by default.
1123 The new counterpiece to "no-xxx" is "enable-xxx".
1125 The patented RC5 and MDC2 algorithms will now be disabled unless
1126 "enable-rc5" and "enable-mdc2", respectively, are specified.
1128 (IDEA remains enabled despite being patented. This is because IDEA
1129 is frequently required for interoperability, and there is no license
1130 fee for non-commercial use. As before, "no-idea" can be used to
1131 avoid this algorithm.)
1135 *) Add processing of proxy certificates (see RFC 3820). This work was
1136 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
1137 EGEE (Enabling Grids for E-science in Europe).
1140 *) RC4 performance overhaul on modern architectures/implementations, such
1141 as Intel P4, IA-64 and AMD64.
1144 *) New utility extract-section.pl. This can be used specify an alternative
1145 section number in a pod file instead of having to treat each file as
1146 a separate case in Makefile. This can be done by adding two lines to the
1149 =for comment openssl_section:XXX
1151 The blank line is mandatory.
1155 *) New arguments -certform, -keyform and -pass for s_client and s_server
1156 to allow alternative format key and certificate files and passphrase
1160 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
1161 update associated structures and add various utility functions.
1163 Add new policy related verify parameters, include policy checking in
1164 standard verify code. Enhance 'smime' application with extra parameters
1165 to support policy checking and print out.
1168 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
1169 Nehemiah processors. These extensions support AES encryption in hardware
1170 as well as RNG (though RNG support is currently disabled).
1171 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
1173 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
1176 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
1177 [Andy Polyakov and a number of other people]
1179 *) Improved PowerPC platform support. Most notably BIGNUM assembler
1180 implementation contributed by IBM.
1181 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
1183 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
1184 exponent rather than 'unsigned long'. There is a corresponding change to
1185 the new 'rsa_keygen' element of the RSA_METHOD structure.
1186 [Jelte Jansen, Geoff Thorpe]
1188 *) Functionality for creating the initial serial number file is now
1189 moved from CA.pl to the 'ca' utility with a new option -create_serial.
1191 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
1192 number file to 1, which is bound to cause problems. To avoid
1193 the problems while respecting compatibility between different 0.9.7
1194 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
1195 CA.pl for serial number initialization. With the new release 0.9.8,
1196 we can fix the problem directly in the 'ca' utility.)
1199 *) Reduced header interdepencies by declaring more opaque objects in
1200 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
1201 give fewer recursive includes, which could break lazy source code - so
1202 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
1203 developers should define this symbol when building and using openssl to
1204 ensure they track the recommended behaviour, interfaces, [etc], but
1205 backwards-compatible behaviour prevails when this isn't defined.
1208 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
1211 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
1212 This will generate a random key of the appropriate length based on the
1213 cipher context. The EVP_CIPHER can provide its own random key generation
1214 routine to support keys of a specific form. This is used in the des and
1215 3des routines to generate a key of the correct parity. Update S/MIME
1216 code to use new functions and hence generate correct parity DES keys.
1217 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
1218 valid (weak or incorrect parity).
1221 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
1222 as looking them up. This is useful when the verified structure may contain
1223 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
1224 present unless the new PKCS7_NO_CRL flag is asserted.
1227 *) Extend ASN1 oid configuration module. It now additionally accepts the
1230 shortName = some long name, 1.2.3.4
1233 *) Reimplemented the BN_CTX implementation. There is now no more static
1234 limitation on the number of variables it can handle nor the depth of the
1235 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
1236 information can now expand as required, and rather than having a single
1237 static array of bignums, BN_CTX now uses a linked-list of such arrays
1238 allowing it to expand on demand whilst maintaining the usefulness of
1239 BN_CTX's "bundling".
1242 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
1243 to allow all RSA operations to function using a single BN_CTX.
1246 *) Preliminary support for certificate policy evaluation and checking. This
1247 is initially intended to pass the tests outlined in "Conformance Testing
1248 of Relying Party Client Certificate Path Processing Logic" v1.07.
1251 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
1252 remained unused and not that useful. A variety of other little bignum
1253 tweaks and fixes have also been made continuing on from the audit (see
1257 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
1258 associated ASN1, EVP and SSL functions and old ASN1 macros.
1261 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
1262 and this should never fail. So the return value from the use of
1263 BN_set_word() (which can fail due to needless expansion) is now deprecated;
1264 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
1267 *) BN_CTX_get() should return zero-valued bignums, providing the same
1268 initialised value as BN_new().
1269 [Geoff Thorpe, suggested by Ulf Möller]
1271 *) Support for inhibitAnyPolicy certificate extension.
1274 *) An audit of the BIGNUM code is underway, for which debugging code is
1275 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
1276 is considered valid when processing BIGNUMs, and causes execution to
1277 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
1278 further steps are taken to deliberately pollute unused data in BIGNUM
1279 structures to try and expose faulty code further on. For now, openssl will
1280 (in its default mode of operation) continue to tolerate the inconsistent
1281 forms that it has tolerated in the past, but authors and packagers should
1282 consider trying openssl and their own applications when compiled with
1283 these debugging symbols defined. It will help highlight potential bugs in
1284 their own code, and will improve the test coverage for OpenSSL itself. At
1285 some point, these tighter rules will become openssl's default to improve
1286 maintainability, though the assert()s and other overheads will remain only
1287 in debugging configurations. See bn.h for more details.
1288 [Geoff Thorpe, Nils Larsch, Ulf Möller]
1290 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
1291 that can only be obtained through BN_CTX_new() (which implicitly
1292 initialises it). The presence of this function only made it possible
1293 to overwrite an existing structure (and cause memory leaks).
1296 *) Because of the callback-based approach for implementing LHASH as a
1297 template type, lh_insert() adds opaque objects to hash-tables and
1298 lh_doall() or lh_doall_arg() are typically used with a destructor callback
1299 to clean up those corresponding objects before destroying the hash table
1300 (and losing the object pointers). So some over-zealous constifications in
1301 LHASH have been relaxed so that lh_insert() does not take (nor store) the
1302 objects as "const" and the lh_doall[_arg] callback wrappers are not
1303 prototyped to have "const" restrictions on the object pointers they are
1304 given (and so aren't required to cast them away any more).
1307 *) The tmdiff.h API was so ugly and minimal that our own timing utility
1308 (speed) prefers to use its own implementation. The two implementations
1309 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
1310 its object type properly exposed (MS_TM) instead of casting to/from "char
1311 *". This may still change yet if someone realises MS_TM and "ms_time_***"
1312 aren't necessarily the greatest nomenclatures - but this is what was used
1313 internally to the implementation so I've used that for now.
1316 *) Ensure that deprecated functions do not get compiled when
1317 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
1318 the self-tests were still using deprecated key-generation functions so
1319 these have been updated also.
1322 *) Reorganise PKCS#7 code to separate the digest location functionality
1323 into PKCS7_find_digest(), digest addtion into PKCS7_bio_add_digest().
1324 New function PKCS7_set_digest() to set the digest type for PKCS#7
1325 digestedData type. Add additional code to correctly generate the
1326 digestedData type and add support for this type in PKCS7 initialization
1330 *) New function PKCS7_set0_type_other() this initializes a PKCS7
1331 structure of type "other".
1334 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
1335 sure the loop does correctly stop and breaking ("division by zero")
1336 modulus operations are not performed. The (pre-generated) prime
1337 table crypto/bn/bn_prime.h was already correct, but it could not be
1338 re-generated on some platforms because of the "division by zero"
1339 situation in the script.
1340 [Ralf S. Engelschall]
1342 *) Update support for ECC-based TLS ciphersuites according to
1343 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
1344 SHA-1 now is only used for "small" curves (where the
1345 representation of a field element takes up to 24 bytes); for
1346 larger curves, the field element resulting from ECDH is directly
1347 used as premaster secret.
1348 [Douglas Stebila (Sun Microsystems Laboratories)]
1350 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
1351 curve secp160r1 to the tests.
1352 [Douglas Stebila (Sun Microsystems Laboratories)]
1354 *) Add the possibility to load symbols globally with DSO.
1355 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
1357 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
1358 control of the error stack.
1361 *) Add support for STORE in ENGINE.
1364 *) Add the STORE type. The intention is to provide a common interface
1365 to certificate and key stores, be they simple file-based stores, or
1366 HSM-type store, or LDAP stores, or...
1367 NOTE: The code is currently UNTESTED and isn't really used anywhere.
1370 *) Add a generic structure called OPENSSL_ITEM. This can be used to
1371 pass a list of arguments to any function as well as provide a way
1372 for a function to pass data back to the caller.
1375 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
1376 works like BUF_strdup() but can be used to duplicate a portion of
1377 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
1381 *) Add the function sk_find_ex() which works like sk_find(), but will
1382 return an index to an element even if an exact match couldn't be
1383 found. The index is guaranteed to point at the element where the
1384 searched-for key would be inserted to preserve sorting order.
1387 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
1388 takes an extra flags argument for optional functionality. Currently,
1389 the following flags are defined:
1391 OBJ_BSEARCH_VALUE_ON_NOMATCH
1392 This one gets OBJ_bsearch_ex() to return a pointer to the first
1393 element where the comparing function returns a negative or zero
1396 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
1397 This one gets OBJ_bsearch_ex() to return a pointer to the first
1398 element where the comparing function returns zero. This is useful
1399 if there are more than one element where the comparing function
1403 *) Make it possible to create self-signed certificates with 'openssl ca'
1404 in such a way that the self-signed certificate becomes part of the
1405 CA database and uses the same mechanisms for serial number generation
1406 as all other certificate signing. The new flag '-selfsign' enables
1407 this functionality. Adapt CA.sh and CA.pl.in.
1410 *) Add functionality to check the public key of a certificate request
1411 against a given private. This is useful to check that a certificate
1412 request can be signed by that key (self-signing).
1415 *) Make it possible to have multiple active certificates with the same
1416 subject in the CA index file. This is done only if the keyword
1417 'unique_subject' is set to 'no' in the main CA section (default
1418 if 'CA_default') of the configuration file. The value is saved
1419 with the database itself in a separate index attribute file,
1420 named like the index file with '.attr' appended to the name.
1423 *) Generate muti valued AVAs using '+' notation in config files for
1427 *) Support for nameConstraints certificate extension.
1430 *) Support for policyConstraints certificate extension.
1433 *) Support for policyMappings certificate extension.
1436 *) Make sure the default DSA_METHOD implementation only uses its
1437 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
1438 and change its own handlers to be NULL so as to remove unnecessary
1439 indirection. This lets alternative implementations fallback to the
1440 default implementation more easily.
1443 *) Support for directoryName in GeneralName related extensions
1447 *) Make it possible to link applications using Makefile.shared.
1448 Make that possible even when linking against static libraries!
1451 *) Support for single pass processing for S/MIME signing. This now
1452 means that S/MIME signing can be done from a pipe, in addition
1453 cleartext signing (multipart/signed type) is effectively streaming
1454 and the signed data does not need to be all held in memory.
1456 This is done with a new flag PKCS7_STREAM. When this flag is set
1457 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
1458 is done after the data is output (and digests calculated) in
1459 SMIME_write_PKCS7().
1462 *) Add full support for -rpath/-R, both in shared libraries and
1463 applications, at least on the platforms where it's known how
1467 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
1468 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
1469 will now compute a table of multiples of the generator that
1470 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
1471 faster (notably in the case of a single point multiplication,
1472 scalar * generator).
1473 [Nils Larsch, Bodo Moeller]
1475 *) IPv6 support for certificate extensions. The various extensions
1476 which use the IP:a.b.c.d can now take IPv6 addresses using the
1477 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
1481 *) Added an ENGINE that implements RSA by performing private key
1482 exponentiations with the GMP library. The conversions to and from
1483 GMP's mpz_t format aren't optimised nor are any montgomery forms
1484 cached, and on x86 it appears OpenSSL's own performance has caught up.
1485 However there are likely to be other architectures where GMP could
1486 provide a boost. This ENGINE is not built in by default, but it can be
1487 specified at Configure time and should be accompanied by the necessary
1488 linker additions, eg;
1489 ./config -DOPENSSL_USE_GMP -lgmp
1492 *) "openssl engine" will not display ENGINE/DSO load failure errors when
1493 testing availability of engines with "-t" - the old behaviour is
1494 produced by increasing the feature's verbosity with "-tt".
1497 *) ECDSA routines: under certain error conditions uninitialized BN objects
1498 could be freed. Solution: make sure initialization is performed early
1499 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
1503 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
1504 and DH_METHOD (eg. by ENGINE implementations) to override the normal
1505 software implementations. For DSA and DH, parameter generation can
1506 also be overriden by providing the appropriate method callbacks.
1509 *) Change the "progress" mechanism used in key-generation and
1510 primality testing to functions that take a new BN_GENCB pointer in
1511 place of callback/argument pairs. The new API functions have "_ex"
1512 postfixes and the older functions are reimplemented as wrappers for
1513 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
1514 declarations of the old functions to help (graceful) attempts to
1515 migrate to the new functions. Also, the new key-generation API
1516 functions operate on a caller-supplied key-structure and return
1517 success/failure rather than returning a key or NULL - this is to
1518 help make "keygen" another member function of RSA_METHOD etc.
1520 Example for using the new callback interface:
1522 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
1526 BN_GENCB_set(&my_cb, my_callback, my_arg);
1528 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
1529 /* For the meaning of a, b in calls to my_callback(), see the
1530 * documentation of the function that calls the callback.
1531 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
1532 * my_callback should return 1 if it wants BN_is_prime_ex()
1533 * to continue, or 0 to stop.
1538 *) Change the ZLIB compression method to be stateful, and make it
1539 available to TLS with the number defined in
1540 draft-ietf-tls-compression-04.txt.
1543 *) Add the ASN.1 structures and functions for CertificatePair, which
1544 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
1546 CertificatePair ::= SEQUENCE {
1547 forward [0] Certificate OPTIONAL,
1548 reverse [1] Certificate OPTIONAL,
1549 -- at least one of the pair shall be present -- }
1551 Also implement the PEM functions to read and write certificate
1552 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
1554 This needed to be defined, mostly for the sake of the LDAP
1555 attribute crossCertificatePair, but may prove useful elsewhere as
1559 *) Make it possible to inhibit symlinking of shared libraries in
1560 Makefile.shared, for Cygwin's sake.
1563 *) Extend the BIGNUM API by creating a function
1564 void BN_set_negative(BIGNUM *a, int neg);
1565 and a macro that behave like
1566 int BN_is_negative(const BIGNUM *a);
1568 to avoid the need to access 'a->neg' directly in applications.
1571 *) Implement fast modular reduction for pseudo-Mersenne primes
1572 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
1573 EC_GROUP_new_curve_GFp() will now automatically use this
1575 [Nils Larsch <nla@trustcenter.de>]
1577 *) Add new lock type (CRYPTO_LOCK_BN).
1580 *) Change the ENGINE framework to automatically load engines
1581 dynamically from specific directories unless they could be
1582 found to already be built in or loaded. Move all the
1583 current engines except for the cryptodev one to a new
1585 The engines in engines/ are built as shared libraries if
1586 the "shared" options was given to ./Configure or ./config.
1587 Otherwise, they are inserted in libcrypto.a.
1588 /usr/local/ssl/engines is the default directory for dynamic
1589 engines, but that can be overriden at configure time through
1590 the usual use of --prefix and/or --openssldir, and at run
1591 time with the environment variable OPENSSL_ENGINES.
1592 [Geoff Thorpe and Richard Levitte]
1594 *) Add Makefile.shared, a helper makefile to build shared
1595 libraries. Addapt Makefile.org.
1598 *) Add version info to Win32 DLLs.
1599 [Peter 'Luna' Runestig" <peter@runestig.com>]
1601 *) Add new 'medium level' PKCS#12 API. Certificates and keys
1602 can be added using this API to created arbitrary PKCS#12
1603 files while avoiding the low level API.
1605 New options to PKCS12_create(), key or cert can be NULL and
1606 will then be omitted from the output file. The encryption
1607 algorithm NIDs can be set to -1 for no encryption, the mac
1608 iteration count can be set to 0 to omit the mac.
1610 Enhance pkcs12 utility by making the -nokeys and -nocerts
1611 options work when creating a PKCS#12 file. New option -nomac
1612 to omit the mac, NONE can be set for an encryption algorithm.
1613 New code is modified to use the enhanced PKCS12_create()
1614 instead of the low level API.
1617 *) Extend ASN1 encoder to support indefinite length constructed
1618 encoding. This can output sequences tags and octet strings in
1619 this form. Modify pk7_asn1.c to support indefinite length
1620 encoding. This is experimental and needs additional code to
1621 be useful, such as an ASN1 bio and some enhanced streaming
1624 Extend template encode functionality so that tagging is passed
1625 down to the template encoder.
1628 *) Let 'openssl req' fail if an argument to '-newkey' is not
1629 recognized instead of using RSA as a default.
1632 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
1633 As these are not official, they are not included in "ALL";
1634 the "ECCdraft" ciphersuite group alias can be used to select them.
1635 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
1637 *) Add ECDH engine support.
1638 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
1640 *) Add ECDH in new directory crypto/ecdh/.
1641 [Douglas Stebila (Sun Microsystems Laboratories)]
1643 *) Let BN_rand_range() abort with an error after 100 iterations
1644 without success (which indicates a broken PRNG).
1647 *) Change BN_mod_sqrt() so that it verifies that the input value
1648 is really the square of the return value. (Previously,
1649 BN_mod_sqrt would show GIGO behaviour.)
1652 *) Add named elliptic curves over binary fields from X9.62, SECG,
1653 and WAP/WTLS; add OIDs that were still missing.
1655 [Sheueling Chang Shantz and Douglas Stebila
1656 (Sun Microsystems Laboratories)]
1658 *) Extend the EC library for elliptic curves over binary fields
1659 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
1662 EC_GF2m_simple_method
1666 EC_GROUP_new_curve_GF2m
1667 EC_GROUP_set_curve_GF2m
1668 EC_GROUP_get_curve_GF2m
1669 EC_POINT_set_affine_coordinates_GF2m
1670 EC_POINT_get_affine_coordinates_GF2m
1671 EC_POINT_set_compressed_coordinates_GF2m
1673 Point compression for binary fields is disabled by default for
1674 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
1677 As binary polynomials are represented as BIGNUMs, various members
1678 of the EC_GROUP and EC_POINT data structures can be shared
1679 between the implementations for prime fields and binary fields;
1680 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
1681 are essentially identical to their ..._GFp counterparts.
1682 (For simplicity, the '..._GFp' prefix has been dropped from
1683 various internal method names.)
1685 An internal 'field_div' method (similar to 'field_mul' and
1686 'field_sqr') has been added; this is used only for binary fields.
1688 [Sheueling Chang Shantz and Douglas Stebila
1689 (Sun Microsystems Laboratories)]
1691 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
1692 through methods ('mul', 'precompute_mult').
1694 The generic implementations (now internally called 'ec_wNAF_mul'
1695 and 'ec_wNAF_precomputed_mult') remain the default if these
1696 methods are undefined.
1698 [Sheueling Chang Shantz and Douglas Stebila
1699 (Sun Microsystems Laboratories)]
1701 *) New function EC_GROUP_get_degree, which is defined through
1702 EC_METHOD. For curves over prime fields, this returns the bit
1703 length of the modulus.
1705 [Sheueling Chang Shantz and Douglas Stebila
1706 (Sun Microsystems Laboratories)]
1708 *) New functions EC_GROUP_dup, EC_POINT_dup.
1709 (These simply call ..._new and ..._copy).
1711 [Sheueling Chang Shantz and Douglas Stebila
1712 (Sun Microsystems Laboratories)]
1714 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
1715 Polynomials are represented as BIGNUMs (where the sign bit is not
1716 used) in the following functions [macros]:
1719 BN_GF2m_sub [= BN_GF2m_add]
1720 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
1721 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
1722 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
1724 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
1725 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
1726 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
1727 BN_GF2m_cmp [= BN_ucmp]
1729 (Note that only the 'mod' functions are actually for fields GF(2^m).
1730 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
1732 For some functions, an the irreducible polynomial defining a
1733 field can be given as an 'unsigned int[]' with strictly
1734 decreasing elements giving the indices of those bits that are set;
1735 i.e., p[] represents the polynomial
1736 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
1738 p[0] > p[1] > ... > p[k] = 0.
1739 This applies to the following functions:
1744 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
1745 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
1747 BN_GF2m_mod_sqrt_arr
1748 BN_GF2m_mod_solve_quad_arr
1752 Conversion can be performed by the following functions:
1757 bntest.c has additional tests for binary polynomial arithmetic.
1759 Two implementations for BN_GF2m_mod_div() are available.
1760 The default algorithm simply uses BN_GF2m_mod_inv() and
1761 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
1762 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
1763 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
1765 [Sheueling Chang Shantz and Douglas Stebila
1766 (Sun Microsystems Laboratories)]
1768 *) Add new error code 'ERR_R_DISABLED' that can be used when some
1769 functionality is disabled at compile-time.
1770 [Douglas Stebila <douglas.stebila@sun.com>]
1772 *) Change default behaviour of 'openssl asn1parse' so that more
1773 information is visible when viewing, e.g., a certificate:
1775 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
1776 mode the content of non-printable OCTET STRINGs is output in a
1777 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
1778 avoid the appearance of a printable string.
1779 [Nils Larsch <nla@trustcenter.de>]
1781 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
1783 EC_GROUP_set_asn1_flag()
1784 EC_GROUP_get_asn1_flag()
1785 EC_GROUP_set_point_conversion_form()
1786 EC_GROUP_get_point_conversion_form()
1787 These control ASN1 encoding details:
1788 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
1789 has been set to OPENSSL_EC_NAMED_CURVE.
1790 - Points are encoded in uncompressed form by default; options for
1791 asn1_for are as for point2oct, namely
1792 POINT_CONVERSION_COMPRESSED
1793 POINT_CONVERSION_UNCOMPRESSED
1794 POINT_CONVERSION_HYBRID
1796 Also add 'seed' and 'seed_len' members to EC_GROUP with access
1799 EC_GROUP_get0_seed()
1800 EC_GROUP_get_seed_len()
1801 This is used only for ASN1 purposes (so far).
1802 [Nils Larsch <nla@trustcenter.de>]
1804 *) Add 'field_type' member to EC_METHOD, which holds the NID
1805 of the appropriate field type OID. The new function
1806 EC_METHOD_get_field_type() returns this value.
1807 [Nils Larsch <nla@trustcenter.de>]
1812 EC_POINT_point2hex()
1813 EC_POINT_hex2point()
1814 providing useful interfaces to EC_POINT_point2oct() and
1815 EC_POINT_oct2point().
1816 [Nils Larsch <nla@trustcenter.de>]
1818 *) Change internals of the EC library so that the functions
1819 EC_GROUP_set_generator()
1820 EC_GROUP_get_generator()
1821 EC_GROUP_get_order()
1822 EC_GROUP_get_cofactor()
1823 are implemented directly in crypto/ec/ec_lib.c and not dispatched
1824 to methods, which would lead to unnecessary code duplication when
1825 adding different types of curves.
1826 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
1828 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
1829 arithmetic, and such that modified wNAFs are generated
1830 (which avoid length expansion in many cases).
1833 *) Add a function EC_GROUP_check_discriminant() (defined via
1834 EC_METHOD) that verifies that the curve discriminant is non-zero.
1836 Add a function EC_GROUP_check() that makes some sanity tests
1837 on a EC_GROUP, its generator and order. This includes
1838 EC_GROUP_check_discriminant().
1839 [Nils Larsch <nla@trustcenter.de>]
1841 *) Add ECDSA in new directory crypto/ecdsa/.
1843 Add applications 'openssl ecparam' and 'openssl ecdsa'
1844 (these are based on 'openssl dsaparam' and 'openssl dsa').
1846 ECDSA support is also included in various other files across the
1847 library. Most notably,
1848 - 'openssl req' now has a '-newkey ecdsa:file' option;
1849 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
1850 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
1851 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
1852 them suitable for ECDSA where domain parameters must be
1853 extracted before the specific public key;
1854 - ECDSA engine support has been added.
1855 [Nils Larsch <nla@trustcenter.de>]
1857 *) Include some named elliptic curves, and add OIDs from X9.62,
1858 SECG, and WAP/WTLS. Each curve can be obtained from the new
1860 EC_GROUP_new_by_curve_name(),
1861 and the list of available named curves can be obtained with
1862 EC_get_builtin_curves().
1863 Also add a 'curve_name' member to EC_GROUP objects, which can be
1865 EC_GROUP_set_curve_name()
1866 EC_GROUP_get_curve_name()
1867 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
1869 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
1870 was actually never needed) and in BN_mul(). The removal in BN_mul()
1871 required a small change in bn_mul_part_recursive() and the addition
1872 of the functions bn_cmp_part_words(), bn_sub_part_words() and
1873 bn_add_part_words(), which do the same thing as bn_cmp_words(),
1874 bn_sub_words() and bn_add_words() except they take arrays with
1878 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
1880 *) Cleanse PEM buffers before freeing them since they may contain
1882 [Benjamin Bennett <ben@psc.edu>]
1884 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
1885 a ciphersuite string such as "DEFAULT:RSA" cannot enable
1886 authentication-only ciphersuites.
1889 *) Since AES128 and AES256 share a single mask bit in the logic of
1890 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
1891 kludge to work properly if AES128 is available and AES256 isn't.
1894 *) Expand security boundary to match 1.1.1 module.
1897 *) Remove redundant features: hash file source, editing of test vectors
1898 modify fipsld to use external fips_premain.c signature.
1901 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
1902 run algorithm test programs.
1905 *) Make algorithm test programs more tolerant of whitespace.
1908 *) Have SSL/TLS server implementation tolerate "mismatched" record
1909 protocol version while receiving ClientHello even if the
1910 ClientHello is fragmented. (The server can't insist on the
1911 particular protocol version it has chosen before the ServerHello
1912 message has informed the client about his choice.)
1915 *) Load error codes if they are not already present instead of using a
1916 static variable. This allows them to be cleanly unloaded and reloaded.
1919 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
1921 *) Introduce limits to prevent malicious keys being able to
1922 cause a denial of service. (CVE-2006-2940)
1923 [Steve Henson, Bodo Moeller]
1925 *) Fix ASN.1 parsing of certain invalid structures that can result
1926 in a denial of service. (CVE-2006-2937) [Steve Henson]
1928 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
1929 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
1931 *) Fix SSL client code which could crash if connecting to a
1932 malicious SSLv2 server. (CVE-2006-4343)
1933 [Tavis Ormandy and Will Drewry, Google Security Team]
1935 *) Change ciphersuite string processing so that an explicit
1936 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
1937 will no longer include "AES128-SHA"), and any other similar
1938 ciphersuite (same bitmap) from *other* protocol versions (so that
1939 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
1940 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
1941 changes from 0.9.8b and 0.9.8d.
1944 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
1946 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
1947 (CVE-2006-4339) [Ben Laurie and Google Security Team]
1949 *) Change the Unix randomness entropy gathering to use poll() when
1950 possible instead of select(), since the latter has some
1951 undesirable limitations.
1952 [Darryl Miles via Richard Levitte and Bodo Moeller]
1954 *) Disable rogue ciphersuites:
1956 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
1957 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
1958 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
1960 The latter two were purportedly from
1961 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
1964 Also deactive the remaining ciphersuites from
1965 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
1966 unofficial, and the ID has long expired.
1969 *) Fix RSA blinding Heisenbug (problems sometimes occured on
1970 dual-core machines) and other potential thread-safety issues.
1973 Changes between 0.9.7i and 0.9.7j [04 May 2006]
1975 *) Adapt fipsld and the build system to link against the validated FIPS
1976 module in FIPS mode.
1979 *) Fixes for VC++ 2005 build under Windows.
1982 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
1983 from a Windows bash shell such as MSYS. It is autodetected from the
1984 "config" script when run from a VC++ environment. Modify standard VC++
1985 build to use fipscanister.o from the GNU make build.
1988 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
1990 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
1991 The value now differs depending on if you build for FIPS or not.
1992 BEWARE! A program linked with a shared FIPSed libcrypto can't be
1993 safely run with a non-FIPSed libcrypto, as it may crash because of
1994 the difference induced by this change.
1997 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
1999 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
2000 (part of SSL_OP_ALL). This option used to disable the
2001 countermeasure against man-in-the-middle protocol-version
2002 rollback in the SSL 2.0 server implementation, which is a bad
2003 idea. (CVE-2005-2969)
2005 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
2006 for Information Security, National Institute of Advanced Industrial
2007 Science and Technology [AIST], Japan)]
2009 *) Minimal support for X9.31 signatures and PSS padding modes. This is
2010 mainly for FIPS compliance and not fully integrated at this stage.
2013 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
2014 the exponentiation using a fixed-length exponent. (Otherwise,
2015 the information leaked through timing could expose the secret key
2016 after many signatures; cf. Bleichenbacher's attack on DSA with
2020 *) Make a new fixed-window mod_exp implementation the default for
2021 RSA, DSA, and DH private-key operations so that the sequence of
2022 squares and multiplies and the memory access pattern are
2023 independent of the particular secret key. This will mitigate
2024 cache-timing and potential related attacks.
2026 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
2027 and this is automatically used by BN_mod_exp_mont() if the new flag
2028 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
2029 will use this BN flag for private exponents unless the flag
2030 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
2031 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
2033 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
2035 *) Change the client implementation for SSLv23_method() and
2036 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
2037 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
2038 (Previously, the SSL 2.0 backwards compatible Client Hello
2039 message format would be used even with SSL_OP_NO_SSLv2.)
2042 *) Add support for smime-type MIME parameter in S/MIME messages which some
2046 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
2047 a threadsafe manner. Modify rsa code to use new function and add calls
2048 to dsa and dh code (which had race conditions before).
2051 *) Include the fixed error library code in the C error file definitions
2052 instead of fixing them up at runtime. This keeps the error code
2053 structures constant.
2056 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
2058 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
2061 *) Fixes for newer kerberos headers. NB: the casts are needed because
2062 the 'length' field is signed on one version and unsigned on another
2063 with no (?) obvious way to tell the difference, without these VC++
2064 complains. Also the "definition" of FAR (blank) is no longer included
2065 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
2066 some needed definitions.
2069 *) Undo Cygwin change.
2072 *) Added support for proxy certificates according to RFC 3820.
2073 Because they may be a security thread to unaware applications,
2074 they must be explicitely allowed in run-time. See
2075 docs/HOWTO/proxy_certificates.txt for further information.
2078 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
2080 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
2081 server and client random values. Previously
2082 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
2083 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
2085 This change has negligible security impact because:
2087 1. Server and client random values still have 24 bytes of pseudo random
2090 2. Server and client random values are sent in the clear in the initial
2093 3. The master secret is derived using the premaster secret (48 bytes in
2094 size for static RSA ciphersuites) as well as client server and random
2097 The OpenSSL team would like to thank the UK NISCC for bringing this issue
2100 [Stephen Henson, reported by UK NISCC]
2102 *) Use Windows randomness collection on Cygwin.
2105 *) Fix hang in EGD/PRNGD query when communication socket is closed
2106 prematurely by EGD/PRNGD.
2107 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
2109 *) Prompt for pass phrases when appropriate for PKCS12 input format.
2112 *) Back-port of selected performance improvements from development
2113 branch, as well as improved support for PowerPC platforms.
2116 *) Add lots of checks for memory allocation failure, error codes to indicate
2117 failure and freeing up memory if a failure occurs.
2118 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
2120 *) Add new -passin argument to dgst.
2123 *) Perform some character comparisons of different types in X509_NAME_cmp:
2124 this is needed for some certificates that reencode DNs into UTF8Strings
2125 (in violation of RFC3280) and can't or wont issue name rollover
2129 *) Make an explicit check during certificate validation to see that
2130 the CA setting in each certificate on the chain is correct. As a
2131 side effect always do the following basic checks on extensions,
2132 not just when there's an associated purpose to the check:
2134 - if there is an unhandled critical extension (unless the user
2135 has chosen to ignore this fault)
2136 - if the path length has been exceeded (if one is set at all)
2137 - that certain extensions fit the associated purpose (if one has
2141 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
2143 *) Avoid a race condition when CRLs are checked in a multi threaded
2144 environment. This would happen due to the reordering of the revoked
2145 entries during signature checking and serial number lookup. Now the
2146 encoding is cached and the serial number sort performed under a lock.
2147 Add new STACK function sk_is_sorted().
2150 *) Add Delta CRL to the extension code.
2153 *) Various fixes to s3_pkt.c so alerts are sent properly.
2154 [David Holmes <d.holmes@f5.com>]
2156 *) Reduce the chances of duplicate issuer name and serial numbers (in
2157 violation of RFC3280) using the OpenSSL certificate creation utilities.
2158 This is done by creating a random 64 bit value for the initial serial
2159 number when a serial number file is created or when a self signed
2160 certificate is created using 'openssl req -x509'. The initial serial
2161 number file is created using 'openssl x509 -next_serial' in CA.pl
2162 rather than being initialized to 1.
2165 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
2167 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
2168 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
2169 [Joe Orton, Steve Henson]
2171 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
2173 [Joe Orton, Steve Henson]
2175 *) Make it possible to have multiple active certificates with the same
2176 subject in the CA index file. This is done only if the keyword
2177 'unique_subject' is set to 'no' in the main CA section (default
2178 if 'CA_default') of the configuration file. The value is saved
2179 with the database itself in a separate index attribute file,
2180 named like the index file with '.attr' appended to the name.
2183 *) X509 verify fixes. Disable broken certificate workarounds when
2184 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
2185 keyUsage extension present. Don't accept CRLs with unhandled critical
2186 extensions: since verify currently doesn't process CRL extensions this
2187 rejects a CRL with *any* critical extensions. Add new verify error codes
2191 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
2192 A clarification of RFC2560 will require the use of OCTET STRINGs and
2193 some implementations cannot handle the current raw format. Since OpenSSL
2194 copies and compares OCSP nonces as opaque blobs without any attempt at
2195 parsing them this should not create any compatibility issues.
2198 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
2199 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
2200 this HMAC (and other) operations are several times slower than OpenSSL
2204 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
2205 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2207 *) Use the correct content when signing type "other".
2210 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
2212 *) Fix various bugs revealed by running the NISCC test suite:
2214 Stop out of bounds reads in the ASN1 code when presented with
2215 invalid tags (CVE-2003-0543 and CVE-2003-0544).
2217 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
2219 If verify callback ignores invalid public key errors don't try to check
2220 certificate signature with the NULL public key.
2224 *) New -ignore_err option in ocsp application to stop the server
2225 exiting on the first error in a request.
2228 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
2229 if the server requested one: as stated in TLS 1.0 and SSL 3.0
2233 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
2234 extra data after the compression methods not only for TLS 1.0
2235 but also for SSL 3.0 (as required by the specification).
2236 [Bodo Moeller; problem pointed out by Matthias Loepfe]
2238 *) Change X509_certificate_type() to mark the key as exported/exportable
2239 when it's 512 *bits* long, not 512 bytes.
2242 *) Change AES_cbc_encrypt() so it outputs exact multiple of
2243 blocks during encryption.
2246 *) Various fixes to base64 BIO and non blocking I/O. On write
2247 flushes were not handled properly if the BIO retried. On read
2248 data was not being buffered properly and had various logic bugs.
2249 This also affects blocking I/O when the data being decoded is a
2253 *) Various S/MIME bugfixes and compatibility changes:
2254 output correct application/pkcs7 MIME type if
2255 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
2256 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
2257 of files as .eml work). Correctly handle very long lines in MIME
2261 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
2263 *) Countermeasure against the Klima-Pokorny-Rosa extension of
2264 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
2265 a protocol version number mismatch like a decryption error
2266 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
2269 *) Turn on RSA blinding by default in the default implementation
2270 to avoid a timing attack. Applications that don't want it can call
2271 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
2272 They would be ill-advised to do so in most cases.
2273 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
2275 *) Change RSA blinding code so that it works when the PRNG is not
2276 seeded (in this case, the secret RSA exponent is abused as
2277 an unpredictable seed -- if it is not unpredictable, there
2278 is no point in blinding anyway). Make RSA blinding thread-safe
2279 by remembering the creator's thread ID in rsa->blinding and
2280 having all other threads use local one-time blinding factors
2281 (this requires more computation than sharing rsa->blinding, but
2282 avoids excessive locking; and if an RSA object is not shared
2283 between threads, blinding will still be very fast).
2286 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
2287 ENGINE as defaults for all supported algorithms irrespective of
2288 the 'flags' parameter. 'flags' is now honoured, so applications
2289 should make sure they are passing it correctly.
2292 *) Target "mingw" now allows native Windows code to be generated in
2293 the Cygwin environment as well as with the MinGW compiler.
2296 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
2298 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
2299 via timing by performing a MAC computation even if incorrrect
2300 block cipher padding has been found. This is a countermeasure
2301 against active attacks where the attacker has to distinguish
2302 between bad padding and a MAC verification error. (CVE-2003-0078)
2304 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
2305 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
2306 Martin Vuagnoux (EPFL, Ilion)]
2308 *) Make the no-err option work as intended. The intention with no-err
2309 is not to have the whole error stack handling routines removed from
2310 libcrypto, it's only intended to remove all the function name and
2311 reason texts, thereby removing some of the footprint that may not
2312 be interesting if those errors aren't displayed anyway.
2314 NOTE: it's still possible for any application or module to have it's
2315 own set of error texts inserted. The routines are there, just not
2316 used by default when no-err is given.
2319 *) Add support for FreeBSD on IA64.
2320 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
2322 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
2323 Kerberos function mit_des_cbc_cksum(). Before this change,
2324 the value returned by DES_cbc_cksum() was like the one from
2325 mit_des_cbc_cksum(), except the bytes were swapped.
2326 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
2328 *) Allow an application to disable the automatic SSL chain building.
2329 Before this a rather primitive chain build was always performed in
2330 ssl3_output_cert_chain(): an application had no way to send the
2331 correct chain if the automatic operation produced an incorrect result.
2333 Now the chain builder is disabled if either:
2335 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
2337 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
2339 The reasoning behind this is that an application would not want the
2340 auto chain building to take place if extra chain certificates are
2341 present and it might also want a means of sending no additional
2342 certificates (for example the chain has two certificates and the
2346 *) Add the possibility to build without the ENGINE framework.
2347 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2349 *) Under Win32 gmtime() can return NULL: check return value in
2350 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
2353 *) DSA routines: under certain error conditions uninitialized BN objects
2354 could be freed. Solution: make sure initialization is performed early
2355 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
2356 Nils Larsch <nla@trustcenter.de> via PR#459)
2359 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
2360 checked on reconnect on the client side, therefore session resumption
2361 could still fail with a "ssl session id is different" error. This
2362 behaviour is masked when SSL_OP_ALL is used due to
2363 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
2364 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2365 followup to PR #377.
2368 *) IA-32 assembler support enhancements: unified ELF targets, support
2369 for SCO/Caldera platforms, fix for Cygwin shared build.
2372 *) Add support for FreeBSD on sparc64. As a consequence, support for
2373 FreeBSD on non-x86 processors is separate from x86 processors on
2374 the config script, much like the NetBSD support.
2375 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
2377 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
2379 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
2382 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
2383 code (06) was taken as the first octet of the session ID and the last
2384 octet was ignored consequently. As a result SSLv2 client side session
2385 caching could not have worked due to the session ID mismatch between
2387 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
2391 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
2392 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
2396 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
2397 seems that in spite of existing for more than a year, many application
2398 author have done nothing to provide the necessary callbacks, which
2399 means that this particular engine will not work properly anywhere.
2400 This is a very unfortunate situation which forces us, in the name
2401 of usability, to give the hw_ncipher.c a static lock, which is part
2403 NOTE: This is for the 0.9.7 series ONLY. This hack will never
2404 appear in 0.9.8 or later. We EXPECT application authors to have
2405 dealt properly with this when 0.9.8 is released (unless we actually
2406 make such changes in the libcrypto locking code that changes will
2407 have to be made anyway).
2410 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
2411 octets have been read, EOF or an error occurs. Without this change
2412 some truncated ASN1 structures will not produce an error.
2415 *) Disable Heimdal support, since it hasn't been fully implemented.
2416 Still give the possibility to force the use of Heimdal, but with
2417 warnings and a request that patches get sent to openssl-dev.
2420 *) Add the VC-CE target, introduce the WINCE sysname, and add
2421 INSTALL.WCE and appropriate conditionals to make it build.
2422 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
2424 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
2425 cygssl-x.y.z.dll, where x, y and z are the major, minor and
2426 edit numbers of the version.
2427 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
2429 *) Introduce safe string copy and catenation functions
2430 (BUF_strlcpy() and BUF_strlcat()).
2431 [Ben Laurie (CHATS) and Richard Levitte]
2433 *) Avoid using fixed-size buffers for one-line DNs.
2434 [Ben Laurie (CHATS)]
2436 *) Add BUF_MEM_grow_clean() to avoid information leakage when
2437 resizing buffers containing secrets, and use where appropriate.
2438 [Ben Laurie (CHATS)]
2440 *) Avoid using fixed size buffers for configuration file location.
2441 [Ben Laurie (CHATS)]
2443 *) Avoid filename truncation for various CA files.
2444 [Ben Laurie (CHATS)]
2446 *) Use sizeof in preference to magic numbers.
2447 [Ben Laurie (CHATS)]
2449 *) Avoid filename truncation in cert requests.
2450 [Ben Laurie (CHATS)]
2452 *) Add assertions to check for (supposedly impossible) buffer
2454 [Ben Laurie (CHATS)]
2456 *) Don't cache truncated DNS entries in the local cache (this could
2457 potentially lead to a spoofing attack).
2458 [Ben Laurie (CHATS)]
2460 *) Fix various buffers to be large enough for hex/decimal
2461 representations in a platform independent manner.
2462 [Ben Laurie (CHATS)]
2464 *) Add CRYPTO_realloc_clean() to avoid information leakage when
2465 resizing buffers containing secrets, and use where appropriate.
2466 [Ben Laurie (CHATS)]
2468 *) Add BIO_indent() to avoid much slightly worrying code to do
2470 [Ben Laurie (CHATS)]
2472 *) Convert sprintf()/BIO_puts() to BIO_printf().
2473 [Ben Laurie (CHATS)]
2475 *) buffer_gets() could terminate with the buffer only half
2477 [Ben Laurie (CHATS)]
2479 *) Add assertions to prevent user-supplied crypto functions from
2480 overflowing internal buffers by having large block sizes, etc.
2481 [Ben Laurie (CHATS)]
2483 *) New OPENSSL_assert() macro (similar to assert(), but enabled
2485 [Ben Laurie (CHATS)]
2487 *) Eliminate unused copy of key in RC4.
2488 [Ben Laurie (CHATS)]
2490 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
2491 [Ben Laurie (CHATS)]
2493 *) Fix off-by-one error in EGD path.
2494 [Ben Laurie (CHATS)]
2496 *) If RANDFILE path is too long, ignore instead of truncating.
2497 [Ben Laurie (CHATS)]
2499 *) Eliminate unused and incorrectly sized X.509 structure
2501 [Ben Laurie (CHATS)]
2503 *) Eliminate unused and dangerous function knumber().
2504 [Ben Laurie (CHATS)]
2506 *) Eliminate unused and dangerous structure, KSSL_ERR.
2507 [Ben Laurie (CHATS)]
2509 *) Protect against overlong session ID context length in an encoded
2510 session object. Since these are local, this does not appear to be
2512 [Ben Laurie (CHATS)]
2514 *) Change from security patch (see 0.9.6e below) that did not affect
2515 the 0.9.6 release series:
2517 Remote buffer overflow in SSL3 protocol - an attacker could
2518 supply an oversized master key in Kerberos-enabled versions.
2520 [Ben Laurie (CHATS)]
2522 *) Change the SSL kerb5 codes to match RFC 2712.
2525 *) Make -nameopt work fully for req and add -reqopt switch.
2526 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
2528 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
2529 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
2531 *) Make sure tests can be performed even if the corresponding algorithms
2532 have been removed entirely. This was also the last step to make
2533 OpenSSL compilable with DJGPP under all reasonable conditions.
2534 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
2536 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
2537 to allow version independent disabling of normally unselected ciphers,
2538 which may be activated as a side-effect of selecting a single cipher.
2540 (E.g., cipher list string "RSA" enables ciphersuites that are left
2541 out of "ALL" because they do not provide symmetric encryption.
2542 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
2543 [Lutz Jaenicke, Bodo Moeller]
2545 *) Add appropriate support for separate platform-dependent build
2546 directories. The recommended way to make a platform-dependent
2547 build directory is the following (tested on Linux), maybe with
2550 # Place yourself outside of the OpenSSL source tree. In
2551 # this example, the environment variable OPENSSL_SOURCE
2552 # is assumed to contain the absolute OpenSSL source directory.
2553 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
2554 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
2555 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
2556 mkdir -p `dirname $F`
2557 ln -s $OPENSSL_SOURCE/$F $F
2560 To be absolutely sure not to disturb the source tree, a "make clean"
2561 is a good thing. If it isn't successfull, don't worry about it,
2562 it probably means the source directory is very clean.
2565 *) Make sure any ENGINE control commands make local copies of string
2566 pointers passed to them whenever necessary. Otherwise it is possible
2567 the caller may have overwritten (or deallocated) the original string
2568 data when a later ENGINE operation tries to use the stored values.
2569 [Götz Babin-Ebell <babinebell@trustcenter.de>]
2571 *) Improve diagnostics in file reading and command-line digests.
2572 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
2574 *) Add AES modes CFB and OFB to the object database. Correct an
2575 error in AES-CFB decryption.
2578 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
2579 allows existing EVP_CIPHER_CTX structures to be reused after
2580 calling EVP_*Final(). This behaviour is used by encryption
2581 BIOs and some applications. This has the side effect that
2582 applications must explicitly clean up cipher contexts with
2583 EVP_CIPHER_CTX_cleanup() or they will leak memory.
2586 *) Check the values of dna and dnb in bn_mul_recursive before calling
2587 bn_mul_comba (a non zero value means the a or b arrays do not contain
2588 n2 elements) and fallback to bn_mul_normal if either is not zero.
2591 *) Fix escaping of non-ASCII characters when using the -subj option
2592 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
2595 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
2596 form for "surname", serialNumber has no short form.
2597 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
2598 therefore remove "mail" short name for "internet 7".
2599 The OID for unique identifiers in X509 certificates is
2600 x500UniqueIdentifier, not uniqueIdentifier.
2601 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
2604 *) Add an "init" command to the ENGINE config module and auto initialize
2605 ENGINEs. Without any "init" command the ENGINE will be initialized
2606 after all ctrl commands have been executed on it. If init=1 the
2607 ENGINE is initailized at that point (ctrls before that point are run
2608 on the uninitialized ENGINE and after on the initialized one). If
2609 init=0 then the ENGINE will not be iniatialized at all.
2612 *) Fix the 'app_verify_callback' interface so that the user-defined
2613 argument is actually passed to the callback: In the
2614 SSL_CTX_set_cert_verify_callback() prototype, the callback
2615 declaration has been changed from
2618 int (*cb)(X509_STORE_CTX *,void *);
2619 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
2620 i=s->ctx->app_verify_callback(&ctx)
2621 has been changed into
2622 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
2624 To update applications using SSL_CTX_set_cert_verify_callback(),
2625 a dummy argument can be added to their callback functions.
2626 [D. K. Smetters <smetters@parc.xerox.com>]
2628 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
2629 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
2631 *) Add and OPENSSL_LOAD_CONF define which will cause
2632 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
2633 This allows older applications to transparently support certain
2634 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
2635 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
2636 load the config file and OPENSSL_add_all_algorithms_conf() which will
2637 always load it have also been added.
2640 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
2641 Adjust NIDs and EVP layer.
2642 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2644 *) Config modules support in openssl utility.
2646 Most commands now load modules from the config file,
2647 though in a few (such as version) this isn't done
2648 because it couldn't be used for anything.
2650 In the case of ca and req the config file used is
2651 the same as the utility itself: that is the -config
2652 command line option can be used to specify an
2656 *) Move default behaviour from OPENSSL_config(). If appname is NULL
2657 use "openssl_conf" if filename is NULL use default openssl config file.
2660 *) Add an argument to OPENSSL_config() to allow the use of an alternative
2661 config section name. Add a new flag to tolerate a missing config file
2662 and move code to CONF_modules_load_file().
2665 *) Support for crypto accelerator cards from Accelerated Encryption
2666 Processing, www.aep.ie. (Use engine 'aep')
2667 The support was copied from 0.9.6c [engine] and adapted/corrected
2668 to work with the new engine framework.
2669 [AEP Inc. and Richard Levitte]
2671 *) Support for SureWare crypto accelerator cards from Baltimore
2672 Technologies. (Use engine 'sureware')
2673 The support was copied from 0.9.6c [engine] and adapted
2674 to work with the new engine framework.
2677 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
2678 make the newer ENGINE framework commands for the CHIL engine work.
2679 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
2681 *) Make it possible to produce shared libraries on ReliantUNIX.
2682 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
2684 *) Add the configuration target debug-linux-ppro.
2685 Make 'openssl rsa' use the general key loading routines
2686 implemented in apps.c, and make those routines able to
2687 handle the key format FORMAT_NETSCAPE and the variant
2689 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2691 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
2692 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
2694 *) Add -keyform to rsautl, and document -engine.
2695 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
2697 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
2698 BIO_R_NO_SUCH_FILE error code rather than the generic
2699 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
2702 *) Add new functions
2704 ERR_peek_last_error_line
2705 ERR_peek_last_error_line_data.
2706 These are similar to
2709 ERR_peek_error_line_data,
2710 but report on the latest error recorded rather than the first one
2711 still in the error queue.
2712 [Ben Laurie, Bodo Moeller]
2714 *) default_algorithms option in ENGINE config module. This allows things
2716 default_algorithms = ALL
2717 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
2720 *) Prelminary ENGINE config module.
2723 *) New experimental application configuration code.
2726 *) Change the AES code to follow the same name structure as all other
2727 symmetric ciphers, and behave the same way. Move everything to
2728 the directory crypto/aes, thereby obsoleting crypto/rijndael.
2729 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
2731 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
2732 [Ben Laurie and Theo de Raadt]
2734 *) Add option to output public keys in req command.
2735 [Massimiliano Pala madwolf@openca.org]
2737 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
2738 (up to about 10% better than before for P-192 and P-224).
2741 *) New functions/macros
2743 SSL_CTX_set_msg_callback(ctx, cb)
2744 SSL_CTX_set_msg_callback_arg(ctx, arg)
2745 SSL_set_msg_callback(ssl, cb)
2746 SSL_set_msg_callback_arg(ssl, arg)
2748 to request calling a callback function
2750 void cb(int write_p, int version, int content_type,
2751 const void *buf, size_t len, SSL *ssl, void *arg)
2753 whenever a protocol message has been completely received
2754 (write_p == 0) or sent (write_p == 1). Here 'version' is the
2755 protocol version according to which the SSL library interprets
2756 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
2757 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
2758 the content type as defined in the SSL 3.0/TLS 1.0 protocol
2759 specification (change_cipher_spec(20), alert(21), handshake(22)).
2760 'buf' and 'len' point to the actual message, 'ssl' to the
2761 SSL object, and 'arg' is the application-defined value set by
2762 SSL[_CTX]_set_msg_callback_arg().
2764 'openssl s_client' and 'openssl s_server' have new '-msg' options
2765 to enable a callback that displays all protocol messages.
2768 *) Change the shared library support so shared libraries are built as
2769 soon as the corresponding static library is finished, and thereby get
2770 openssl and the test programs linked against the shared library.
2771 This still only happens when the keyword "shard" has been given to
2772 the configuration scripts.
2774 NOTE: shared library support is still an experimental thing, and
2775 backward binary compatibility is still not guaranteed.
2776 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
2778 *) Add support for Subject Information Access extension.
2779 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
2781 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
2782 additional bytes when new memory had to be allocated, not just
2783 when reusing an existing buffer.
2786 *) New command line and configuration option 'utf8' for the req command.
2787 This allows field values to be specified as UTF8 strings.
2790 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
2791 runs for the former and machine-readable output for the latter.
2794 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
2795 of the e-mail address in the DN (i.e., it will go into a certificate
2796 extension only). The new configuration file option 'email_in_dn = no'
2797 has the same effect.
2798 [Massimiliano Pala madwolf@openca.org]
2800 *) Change all functions with names starting with des_ to be starting
2801 with DES_ instead. Add wrappers that are compatible with libdes,
2802 but are named _ossl_old_des_*. Finally, add macros that map the
2803 des_* symbols to the corresponding _ossl_old_des_* if libdes
2804 compatibility is desired. If OpenSSL 0.9.6c compatibility is
2805 desired, the des_* symbols will be mapped to DES_*, with one
2808 Since we provide two compatibility mappings, the user needs to
2809 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
2810 compatibility is desired. The default (i.e., when that macro
2811 isn't defined) is OpenSSL 0.9.6c compatibility.
2813 There are also macros that enable and disable the support of old
2814 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
2815 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
2816 are defined, the default will apply: to support the old des routines.
2818 In either case, one must include openssl/des.h to get the correct
2819 definitions. Do not try to just include openssl/des_old.h, that
2822 NOTE: This is a major break of an old API into a new one. Software
2823 authors are encouraged to switch to the DES_ style functions. Some
2824 time in the future, des_old.h and the libdes compatibility functions
2825 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
2826 default), and then completely removed.
2829 *) Test for certificates which contain unsupported critical extensions.
2830 If such a certificate is found during a verify operation it is
2831 rejected by default: this behaviour can be overridden by either
2832 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
2833 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
2834 X509_supported_extension() has also been added which returns 1 if a
2835 particular extension is supported.
2838 *) Modify the behaviour of EVP cipher functions in similar way to digests
2839 to retain compatibility with existing code.
2842 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
2843 compatibility with existing code. In particular the 'ctx' parameter does
2844 not have to be to be initialized before the call to EVP_DigestInit() and
2845 it is tidied up after a call to EVP_DigestFinal(). New function
2846 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
2847 EVP_MD_CTX_copy() changed to not require the destination to be
2848 initialized valid and new function EVP_MD_CTX_copy_ex() added which
2849 requires the destination to be valid.
2851 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
2852 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
2855 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
2856 so that complete 'Handshake' protocol structures are kept in memory
2857 instead of overwriting 'msg_type' and 'length' with 'body' data.
2860 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
2861 [Massimo Santin via Richard Levitte]
2863 *) Major restructuring to the underlying ENGINE code. This includes
2864 reduction of linker bloat, separation of pure "ENGINE" manipulation
2865 (initialisation, etc) from functionality dealing with implementations
2866 of specific crypto iterfaces. This change also introduces integrated
2867 support for symmetric ciphers and digest implementations - so ENGINEs
2868 can now accelerate these by providing EVP_CIPHER and EVP_MD
2869 implementations of their own. This is detailed in crypto/engine/README
2870 as it couldn't be adequately described here. However, there are a few
2871 API changes worth noting - some RSA, DSA, DH, and RAND functions that
2872 were changed in the original introduction of ENGINE code have now
2873 reverted back - the hooking from this code to ENGINE is now a good
2874 deal more passive and at run-time, operations deal directly with
2875 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
2876 dereferencing through an ENGINE pointer any more. Also, the ENGINE
2877 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
2878 they were not being used by the framework as there is no concept of a
2879 BIGNUM_METHOD and they could not be generalised to the new
2880 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
2881 ENGINE_cpy() has been removed as it cannot be consistently defined in
2885 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
2888 *) Change mkdef.pl to sort symbols that get the same entry number,
2889 and make sure the automatically generated functions ERR_load_*
2890 become part of libeay.num as well.
2893 *) New function SSL_renegotiate_pending(). This returns true once
2894 renegotiation has been requested (either SSL_renegotiate() call
2895 or HelloRequest/ClientHello receveived from the peer) and becomes
2896 false once a handshake has been completed.
2897 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
2898 sends a HelloRequest, but does not ensure that a handshake takes
2899 place. SSL_renegotiate_pending() is useful for checking if the
2900 client has followed the request.)
2903 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
2904 By default, clients may request session resumption even during
2905 renegotiation (if session ID contexts permit); with this option,
2906 session resumption is possible only in the first handshake.
2908 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
2909 more bits available for options that should not be part of
2910 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
2913 *) Add some demos for certificate and certificate request creation.
2916 *) Make maximum certificate chain size accepted from the peer application
2917 settable (SSL*_get/set_max_cert_list()), as proposed by
2918 "Douglas E. Engert" <deengert@anl.gov>.
2921 *) Add support for shared libraries for Unixware-7
2922 (Boyd Lynn Gerber <gerberb@zenez.com>).
2925 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
2926 be done prior to destruction. Use this to unload error strings from
2927 ENGINEs that load their own error strings. NB: This adds two new API
2928 functions to "get" and "set" this destroy handler in an ENGINE.
2931 *) Alter all existing ENGINE implementations (except "openssl" and
2932 "openbsd") to dynamically instantiate their own error strings. This
2933 makes them more flexible to be built both as statically-linked ENGINEs
2934 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
2935 Also, add stub code to each that makes building them as self-contained
2936 shared-libraries easier (see README.ENGINE).
2939 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
2940 implementations into applications that are completely implemented in
2941 self-contained shared-libraries. The "dynamic" ENGINE exposes control
2942 commands that can be used to configure what shared-library to load and
2943 to control aspects of the way it is handled. Also, made an update to
2944 the README.ENGINE file that brings its information up-to-date and
2945 provides some information and instructions on the "dynamic" ENGINE
2946 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
2949 *) Make it possible to unload ranges of ERR strings with a new
2950 "ERR_unload_strings" function.
2953 *) Add a copy() function to EVP_MD.
2956 *) Make EVP_MD routines take a context pointer instead of just the
2957 md_data void pointer.
2960 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
2961 that the digest can only process a single chunk of data
2962 (typically because it is provided by a piece of
2963 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
2964 is only going to provide a single chunk of data, and hence the
2965 framework needn't accumulate the data for oneshot drivers.
2968 *) As with "ERR", make it possible to replace the underlying "ex_data"
2969 functions. This change also alters the storage and management of global
2970 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
2971 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
2972 index counters. The API functions that use this state have been changed
2973 to take a "class_index" rather than pointers to the class's local STACK
2974 and counter, and there is now an API function to dynamically create new
2975 classes. This centralisation allows us to (a) plug a lot of the
2976 thread-safety problems that existed, and (b) makes it possible to clean
2977 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
2978 such data would previously have always leaked in application code and
2979 workarounds were in place to make the memory debugging turn a blind eye
2980 to it. Application code that doesn't use this new function will still
2981 leak as before, but their memory debugging output will announce it now
2982 rather than letting it slide.
2984 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
2985 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
2986 has a return value to indicate success or failure.
2989 *) Make it possible to replace the underlying "ERR" functions such that the
2990 global state (2 LHASH tables and 2 locks) is only used by the "default"
2991 implementation. This change also adds two functions to "get" and "set"
2992 the implementation prior to it being automatically set the first time
2993 any other ERR function takes place. Ie. an application can call "get",
2994 pass the return value to a module it has just loaded, and that module
2995 can call its own "set" function using that value. This means the
2996 module's "ERR" operations will use (and modify) the error state in the
2997 application and not in its own statically linked copy of OpenSSL code.
3000 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
3001 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
3002 the operation, and provides a more encapsulated way for external code
3003 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
3004 to use these functions rather than manually incrementing the counts.
3006 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
3009 *) Add EVP test program.
3012 *) Add symmetric cipher support to ENGINE. Expect the API to change!
3015 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
3016 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
3017 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
3018 These allow a CRL to be built without having to access X509_CRL fields
3019 directly. Modify 'ca' application to use new functions.
3022 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
3023 bug workarounds. Rollback attack detection is a security feature.
3024 The problem will only arise on OpenSSL servers when TLSv1 is not
3025 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
3026 Software authors not wanting to support TLSv1 will have special reasons
3027 for their choice and can explicitly enable this option.
3028 [Bodo Moeller, Lutz Jaenicke]
3030 *) Rationalise EVP so it can be extended: don't include a union of
3031 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
3032 (similar to those existing for EVP_CIPHER_CTX).
3037 EVP_MD_CTX_init(&md); /* new function call */
3038 EVP_DigestInit(&md, EVP_sha1());
3039 EVP_DigestUpdate(&md, in, len);
3040 EVP_DigestFinal(&md, out, NULL);
3041 EVP_MD_CTX_cleanup(&md); /* new function call */
3045 *) Make DES key schedule conform to the usual scheme, as well as
3046 correcting its structure. This means that calls to DES functions
3047 now have to pass a pointer to a des_key_schedule instead of a
3048 plain des_key_schedule (which was actually always a pointer
3051 des_key_schedule ks;
3053 des_set_key_checked(..., &ks);
3054 des_ncbc_encrypt(..., &ks, ...);
3056 (Note that a later change renames 'des_...' into 'DES_...'.)
3059 *) Initial reduction of linker bloat: the use of some functions, such as
3060 PEM causes large amounts of unused functions to be linked in due to
3061 poor organisation. For example pem_all.c contains every PEM function
3062 which has a knock on effect of linking in large amounts of (unused)
3063 ASN1 code. Grouping together similar functions and splitting unrelated
3064 functions prevents this.
3067 *) Cleanup of EVP macros.
3070 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
3071 correct _ecb suffix.
3074 *) Add initial OCSP responder support to ocsp application. The
3075 revocation information is handled using the text based index
3076 use by the ca application. The responder can either handle
3077 requests generated internally, supplied in files (for example
3078 via a CGI script) or using an internal minimal server.
3081 *) Add configuration choices to get zlib compression for TLS.
3084 *) Changes to Kerberos SSL for RFC 2712 compliance:
3085 1. Implemented real KerberosWrapper, instead of just using
3086 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
3087 2. Implemented optional authenticator field of KerberosWrapper.
3089 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
3090 and authenticator structs; see crypto/krb5/.
3092 Generalized Kerberos calls to support multiple Kerberos libraries.
3093 [Vern Staats <staatsvr@asc.hpc.mil>,
3094 Jeffrey Altman <jaltman@columbia.edu>
3095 via Richard Levitte]
3097 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
3098 already does with RSA. testdsa.h now has 'priv_key/pub_key'
3099 values for each of the key sizes rather than having just
3100 parameters (and 'speed' generating keys each time).
3103 *) Speed up EVP routines.
3106 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
3107 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
3108 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
3109 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
3111 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
3112 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
3113 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
3116 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
3118 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
3121 *) Added the OS2-EMX target.
3122 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
3124 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
3125 to support NCONF routines in extension code. New function CONF_set_nconf()
3126 to allow functions which take an NCONF to also handle the old LHASH
3127 structure: this means that the old CONF compatible routines can be
3128 retained (in particular wrt extensions) without having to duplicate the
3129 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
3132 *) Enhance the general user interface with mechanisms for inner control
3133 and with possibilities to have yes/no kind of prompts.
3136 *) Change all calls to low level digest routines in the library and
3137 applications to use EVP. Add missing calls to HMAC_cleanup() and
3138 don't assume HMAC_CTX can be copied using memcpy().
3139 [Verdon Walker <VWalker@novell.com>, Steve Henson]
3141 *) Add the possibility to control engines through control names but with
3142 arbitrary arguments instead of just a string.
3143 Change the key loaders to take a UI_METHOD instead of a callback
3144 function pointer. NOTE: this breaks binary compatibility with earlier
3145 versions of OpenSSL [engine].
3146 Adapt the nCipher code for these new conditions and add a card insertion
3150 *) Enhance the general user interface with mechanisms to better support
3151 dialog box interfaces, application-defined prompts, the possibility
3152 to use defaults (for example default passwords from somewhere else)
3153 and interrupts/cancellations.
3156 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
3157 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
3160 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
3161 tidy up some unnecessarily weird code in 'sk_new()').
3162 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
3164 *) Change the key loading routines for ENGINEs to use the same kind
3165 callback (pem_password_cb) as all other routines that need this
3169 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
3170 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
3171 than this minimum value is recommended.
3174 *) New random seeder for OpenVMS, using the system process statistics
3175 that are easily reachable.
3178 *) Windows apparently can't transparently handle global
3179 variables defined in DLLs. Initialisations such as:
3181 const ASN1_ITEM *it = &ASN1_INTEGER_it;
3183 wont compile. This is used by the any applications that need to
3184 declare their own ASN1 modules. This was fixed by adding the option
3185 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
3186 needed for static libraries under Win32.
3189 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
3190 setting of purpose and trust fields. New X509_STORE trust and
3191 purpose functions and tidy up setting in other SSL functions.
3194 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
3195 structure. These are inherited by X509_STORE_CTX when it is
3196 initialised. This allows various defaults to be set in the
3197 X509_STORE structure (such as flags for CRL checking and custom
3198 purpose or trust settings) for functions which only use X509_STORE_CTX
3199 internally such as S/MIME.
3201 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
3202 trust settings if they are not set in X509_STORE. This allows X509_STORE
3203 purposes and trust (in S/MIME for example) to override any set by default.
3205 Add command line options for CRL checking to smime, s_client and s_server
3209 *) Initial CRL based revocation checking. If the CRL checking flag(s)
3210 are set then the CRL is looked up in the X509_STORE structure and
3211 its validity and signature checked, then if the certificate is found
3212 in the CRL the verify fails with a revoked error.
3214 Various new CRL related callbacks added to X509_STORE_CTX structure.
3216 Command line options added to 'verify' application to support this.
3218 This needs some additional work, such as being able to handle multiple
3219 CRLs with different times, extension based lookup (rather than just
3220 by subject name) and ultimately more complete V2 CRL extension
3224 *) Add a general user interface API (crypto/ui/). This is designed
3225 to replace things like des_read_password and friends (backward
3226 compatibility functions using this new API are provided).
3227 The purpose is to remove prompting functions from the DES code
3228 section as well as provide for prompting through dialog boxes in
3229 a window system and the like.
3232 *) Add "ex_data" support to ENGINE so implementations can add state at a
3233 per-structure level rather than having to store it globally.
3236 *) Make it possible for ENGINE structures to be copied when retrieved by
3237 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
3238 This causes the "original" ENGINE structure to act like a template,
3239 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
3240 operational state can be localised to each ENGINE structure, despite the
3241 fact they all share the same "methods". New ENGINE structures returned in
3242 this case have no functional references and the return value is the single
3243 structural reference. This matches the single structural reference returned
3244 by ENGINE_by_id() normally, when it is incremented on the pre-existing
3248 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
3249 needs to match any other type at all we need to manually clear the
3253 *) Changes to the "openssl engine" utility to include;
3254 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
3255 about an ENGINE's available control commands.
3256 - executing control commands from command line arguments using the
3257 '-pre' and '-post' switches. '-post' is only used if '-t' is
3258 specified and the ENGINE is successfully initialised. The syntax for
3259 the individual commands are colon-separated, for example;
3260 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
3263 *) New dynamic control command support for ENGINEs. ENGINEs can now
3264 declare their own commands (numbers), names (strings), descriptions,
3265 and input types for run-time discovery by calling applications. A
3266 subset of these commands are implicitly classed as "executable"
3267 depending on their input type, and only these can be invoked through
3268 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
3269 can be based on user input, config files, etc). The distinction is
3270 that "executable" commands cannot return anything other than a boolean
3271 result and can only support numeric or string input, whereas some
3272 discoverable commands may only be for direct use through
3273 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
3274 pointers, or other custom uses. The "executable" commands are to
3275 support parameterisations of ENGINE behaviour that can be
3276 unambiguously defined by ENGINEs and used consistently across any
3277 OpenSSL-based application. Commands have been added to all the
3278 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
3279 control over shared-library paths without source code alterations.
3282 *) Changed all ENGINE implementations to dynamically allocate their
3283 ENGINEs rather than declaring them statically. Apart from this being
3284 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
3285 this also allows the implementations to compile without using the
3286 internal engine_int.h header.
3289 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
3290 'const' value. Any code that should be able to modify a RAND_METHOD
3291 should already have non-const pointers to it (ie. they should only
3292 modify their own ones).
3295 *) Made a variety of little tweaks to the ENGINE code.
3296 - "atalla" and "ubsec" string definitions were moved from header files
3297 to C code. "nuron" string definitions were placed in variables
3298 rather than hard-coded - allowing parameterisation of these values
3299 later on via ctrl() commands.
3300 - Removed unused "#if 0"'d code.
3301 - Fixed engine list iteration code so it uses ENGINE_free() to release
3302 structural references.
3303 - Constified the RAND_METHOD element of ENGINE structures.
3304 - Constified various get/set functions as appropriate and added
3305 missing functions (including a catch-all ENGINE_cpy that duplicates
3306 all ENGINE values onto a new ENGINE except reference counts/state).
3307 - Removed NULL parameter checks in get/set functions. Setting a method
3308 or function to NULL is a way of cancelling out a previously set
3309 value. Passing a NULL ENGINE parameter is just plain stupid anyway
3310 and doesn't justify the extra error symbols and code.
3311 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
3312 flags from engine_int.h to engine.h.
3313 - Changed prototypes for ENGINE handler functions (init(), finish(),
3314 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
3317 *) Implement binary inversion algorithm for BN_mod_inverse in addition
3318 to the algorithm using long division. The binary algorithm can be
3319 used only if the modulus is odd. On 32-bit systems, it is faster
3320 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
3321 roughly 5-15% for 256-bit moduli), so we use it only for moduli
3322 up to 450 bits. In 64-bit environments, the binary algorithm
3323 appears to be advantageous for much longer moduli; here we use it
3324 for moduli up to 2048 bits.
3327 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
3328 could not support the combine flag in choice fields.
3331 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
3332 extensions from a certificate request to the certificate.
3335 *) Allow multiple 'certopt' and 'nameopt' options to be separated
3336 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
3337 file: this allows the display of the certificate about to be
3338 signed to be customised, to allow certain fields to be included
3339 or excluded and extension details. The old system didn't display
3340 multicharacter strings properly, omitted fields not in the policy
3341 and couldn't display additional details such as extensions.
3344 *) Function EC_POINTs_mul for multiple scalar multiplication
3345 of an arbitrary number of elliptic curve points
3346 \sum scalars[i]*points[i],
3347 optionally including the generator defined for the EC_GROUP:
3348 scalar*generator + \sum scalars[i]*points[i].
3350 EC_POINT_mul is a simple wrapper function for the typical case
3351 that the point list has just one item (besides the optional
3355 *) First EC_METHODs for curves over GF(p):
3357 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
3358 operations and provides various method functions that can also
3359 operate with faster implementations of modular arithmetic.
3361 EC_GFp_mont_method() reuses most functions that are part of
3362 EC_GFp_simple_method, but uses Montgomery arithmetic.
3364 [Bodo Moeller; point addition and point doubling
3365 implementation directly derived from source code provided by
3366 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
3368 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
3369 crypto/ec/ec_lib.c):
3371 Curves are EC_GROUP objects (with an optional group generator)
3372 based on EC_METHODs that are built into the library.
3374 Points are EC_POINT objects based on EC_GROUP objects.
3376 Most of the framework would be able to handle curves over arbitrary
3377 finite fields, but as there are no obvious types for fields other
3378 than GF(p), some functions are limited to that for now.
3381 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
3382 that the file contains a complete HTTP response.
3385 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
3386 change the def and num file printf format specifier from "%-40sXXX"
3387 to "%-39s XXX". The latter will always guarantee a space after the
3388 field while the former will cause them to run together if the field
3389 is 40 of more characters long.
3392 *) Constify the cipher and digest 'method' functions and structures
3393 and modify related functions to take constant EVP_MD and EVP_CIPHER
3397 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
3398 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
3401 *) Modify EVP_Digest*() routines so they now return values. Although the
3402 internal software routines can never fail additional hardware versions
3406 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
3408 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
3409 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
3412 ERR_R_NESTED_ASN1_ERROR
3414 ERR_R_MISSING_ASN1_EOS
3415 were 4 .. 9, conflicting with
3416 ERR_LIB_RSA (= ERR_R_RSA_LIB)
3418 ERR_LIB_PEM (= ERR_R_PEM_LIB).
3419 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
3421 Add new error code 'ERR_R_INTERNAL_ERROR'.
3424 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
3428 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
3429 sets the subject name for a new request or supersedes the
3430 subject name in a given request. Formats that can be parsed are
3431 'CN=Some Name, OU=myOU, C=IT'
3433 'CN=Some Name/OU=myOU/C=IT'.
3435 Add options '-batch' and '-verbose' to 'openssl req'.
3436 [Massimiliano Pala <madwolf@hackmasters.net>]
3438 *) Introduce the possibility to access global variables through
3439 functions on platform were that's the best way to handle exporting
3440 global variables in shared libraries. To enable this functionality,
3441 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
3442 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
3443 is normally done by Configure or something similar).
3445 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
3446 in the source file (foo.c) like this:
3448 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
3449 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
3451 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
3452 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
3454 OPENSSL_DECLARE_GLOBAL(int,foo);
3455 #define foo OPENSSL_GLOBAL_REF(foo)
3456 OPENSSL_DECLARE_GLOBAL(double,bar);
3457 #define bar OPENSSL_GLOBAL_REF(bar)
3459 The #defines are very important, and therefore so is including the
3460 header file everywhere where the defined globals are used.
3462 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
3463 of ASN.1 items, but that structure is a bit different.
3465 The largest change is in util/mkdef.pl which has been enhanced with
3466 better and easier to understand logic to choose which symbols should
3467 go into the Windows .def files as well as a number of fixes and code
3468 cleanup (among others, algorithm keywords are now sorted
3469 lexicographically to avoid constant rewrites).
3472 *) In BN_div() keep a copy of the sign of 'num' before writing the
3473 result to 'rm' because if rm==num the value will be overwritten
3474 and produce the wrong result if 'num' is negative: this caused
3475 problems with BN_mod() and BN_nnmod().
3478 *) Function OCSP_request_verify(). This checks the signature on an
3479 OCSP request and verifies the signer certificate. The signer
3480 certificate is just checked for a generic purpose and OCSP request
3484 *) Add OCSP_check_validity() function to check the validity of OCSP
3485 responses. OCSP responses are prepared in real time and may only
3486 be a few seconds old. Simply checking that the current time lies
3487 between thisUpdate and nextUpdate max reject otherwise valid responses
3488 caused by either OCSP responder or client clock inaccuracy. Instead
3489 we allow thisUpdate and nextUpdate to fall within a certain period of
3490 the current time. The age of the response can also optionally be
3491 checked. Two new options -validity_period and -status_age added to
3495 *) If signature or public key algorithm is unrecognized print out its
3496 OID rather that just UNKNOWN.
3499 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
3500 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
3501 ID to be generated from the issuer certificate alone which can then be
3502 passed to OCSP_id_issuer_cmp().
3505 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
3506 ASN1 modules to export functions returning ASN1_ITEM pointers
3507 instead of the ASN1_ITEM structures themselves. This adds several
3508 new macros which allow the underlying ASN1 function/structure to
3509 be accessed transparently. As a result code should not use ASN1_ITEM
3510 references directly (such as &X509_it) but instead use the relevant
3511 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
3512 use of the new ASN1 code on platforms where exporting structures
3513 is problematical (for example in shared libraries) but exporting
3514 functions returning pointers to structures is not.
3517 *) Add support for overriding the generation of SSL/TLS session IDs.
3518 These callbacks can be registered either in an SSL_CTX or per SSL.
3519 The purpose of this is to allow applications to control, if they wish,
3520 the arbitrary values chosen for use as session IDs, particularly as it
3521 can be useful for session caching in multiple-server environments. A
3522 command-line switch for testing this (and any client code that wishes
3523 to use such a feature) has been added to "s_server".
3524 [Geoff Thorpe, Lutz Jaenicke]
3526 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
3527 of the form '#if defined(...) || defined(...) || ...' and
3528 '#if !defined(...) && !defined(...) && ...'. This also avoids
3529 the growing number of special cases it was previously handling.
3532 *) Make all configuration macros available for application by making
3533 sure they are available in opensslconf.h, by giving them names starting
3534 with "OPENSSL_" to avoid conflicts with other packages and by making
3535 sure e_os2.h will cover all platform-specific cases together with
3537 Additionally, it is now possible to define configuration/platform-
3538 specific names (called "system identities"). In the C code, these
3539 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
3540 macro with the name beginning with "OPENSSL_SYS_", which is determined
3541 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
3545 *) New option -set_serial to 'req' and 'x509' this allows the serial
3546 number to use to be specified on the command line. Previously self
3547 signed certificates were hard coded with serial number 0 and the
3548 CA options of 'x509' had to use a serial number in a file which was
3552 *) New options to 'ca' utility to support V2 CRL entry extensions.
3553 Currently CRL reason, invalidity date and hold instruction are
3554 supported. Add new CRL extensions to V3 code and some new objects.
3557 *) New function EVP_CIPHER_CTX_set_padding() this is used to
3558 disable standard block padding (aka PKCS#5 padding) in the EVP
3559 API, which was previously mandatory. This means that the data is
3560 not padded in any way and so the total length much be a multiple
3561 of the block size, otherwise an error occurs.
3564 *) Initial (incomplete) OCSP SSL support.
3567 *) New function OCSP_parse_url(). This splits up a URL into its host,
3568 port and path components: primarily to parse OCSP URLs. New -url
3569 option to ocsp utility.
3572 *) New nonce behavior. The return value of OCSP_check_nonce() now
3573 reflects the various checks performed. Applications can decide
3574 whether to tolerate certain situations such as an absent nonce
3575 in a response when one was present in a request: the ocsp application
3576 just prints out a warning. New function OCSP_add1_basic_nonce()
3577 this is to allow responders to include a nonce in a response even if
3578 the request is nonce-less.
3581 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
3582 skipped when using openssl x509 multiple times on a single input file,
3583 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
3586 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
3587 set string type: to handle setting ASN1_TIME structures. Fix ca
3588 utility to correctly initialize revocation date of CRLs.
3591 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
3592 the clients preferred ciphersuites and rather use its own preferences.
3593 Should help to work around M$ SGC (Server Gated Cryptography) bug in
3594 Internet Explorer by ensuring unchanged hash method during stepup.
3595 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
3598 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
3599 to aes and add a new 'exist' option to print out symbols that don't
3603 *) Additional options to ocsp utility to allow flags to be set and
3604 additional certificates supplied.
3607 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
3608 OCSP client a number of certificate to only verify the response
3612 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
3613 handle the new API. Currently only ECB, CBC modes supported. Add new
3616 Add TLS AES ciphersuites as described in RFC3268, "Advanced
3617 Encryption Standard (AES) Ciphersuites for Transport Layer
3618 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
3619 not enabled by default and were not part of the "ALL" ciphersuite
3620 alias because they were not yet official; they could be
3621 explicitly requested by specifying the "AESdraft" ciphersuite
3622 group alias. In the final release of OpenSSL 0.9.7, the group
3623 alias is called "AES" and is part of "ALL".)
3624 [Ben Laurie, Steve Henson, Bodo Moeller]
3626 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
3627 request to response.
3630 *) Functions for OCSP responders. OCSP_request_onereq_count(),
3631 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
3632 extract information from a certificate request. OCSP_response_create()
3633 creates a response and optionally adds a basic response structure.
3634 OCSP_basic_add1_status() adds a complete single response to a basic
3635 response and returns the OCSP_SINGLERESP structure just added (to allow
3636 extensions to be included for example). OCSP_basic_add1_cert() adds a
3637 certificate to a basic response and OCSP_basic_sign() signs a basic
3638 response with various flags. New helper functions ASN1_TIME_check()
3639 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
3640 (converts ASN1_TIME to GeneralizedTime).
3643 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
3644 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
3645 structure from a certificate. X509_pubkey_digest() digests the public_key
3646 contents: this is used in various key identifiers.
3649 *) Make sk_sort() tolerate a NULL argument.
3650 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
3652 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
3653 passed by the function are trusted implicitly. If any of them signed the
3654 response then it is assumed to be valid and is not verified.
3657 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
3658 to data. This was previously part of the PKCS7 ASN1 code. This
3659 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
3660 [Steve Henson, reported by Kenneth R. Robinette
3661 <support@securenetterm.com>]
3663 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
3664 routines: without these tracing memory leaks is very painful.
3665 Fix leaks in PKCS12 and PKCS7 routines.
3668 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
3669 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
3670 effectively meant GeneralizedTime would never be used. Now it
3671 is initialised to -1 but X509_time_adj() now has to check the value
3672 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
3673 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
3674 [Steve Henson, reported by Kenneth R. Robinette
3675 <support@securenetterm.com>]
3677 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
3678 result in a zero length in the ASN1_INTEGER structure which was
3679 not consistent with the structure when d2i_ASN1_INTEGER() was used
3680 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
3681 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
3682 where it did not print out a minus for negative ASN1_INTEGER.
3685 *) Add summary printout to ocsp utility. The various functions which
3686 convert status values to strings have been renamed to:
3687 OCSP_response_status_str(), OCSP_cert_status_str() and
3688 OCSP_crl_reason_str() and are no longer static. New options
3689 to verify nonce values and to disable verification. OCSP response
3690 printout format cleaned up.
3693 *) Add additional OCSP certificate checks. These are those specified
3694 in RFC2560. This consists of two separate checks: the CA of the
3695 certificate being checked must either be the OCSP signer certificate
3696 or the issuer of the OCSP signer certificate. In the latter case the
3697 OCSP signer certificate must contain the OCSP signing extended key
3698 usage. This check is performed by attempting to match the OCSP
3699 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
3700 in the OCSP_CERTID structures of the response.
3703 *) Initial OCSP certificate verification added to OCSP_basic_verify()
3704 and related routines. This uses the standard OpenSSL certificate
3705 verify routines to perform initial checks (just CA validity) and
3706 to obtain the certificate chain. Then additional checks will be
3707 performed on the chain. Currently the root CA is checked to see
3708 if it is explicitly trusted for OCSP signing. This is used to set
3709 a root CA as a global signing root: that is any certificate that
3710 chains to that CA is an acceptable OCSP signing certificate.
3713 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
3714 extensions from a separate configuration file.
3715 As when reading extensions from the main configuration file,
3716 the '-extensions ...' option may be used for specifying the
3718 [Massimiliano Pala <madwolf@comune.modena.it>]
3720 *) New OCSP utility. Allows OCSP requests to be generated or
3721 read. The request can be sent to a responder and the output
3722 parsed, outputed or printed in text form. Not complete yet:
3723 still needs to check the OCSP response validity.
3726 *) New subcommands for 'openssl ca':
3727 'openssl ca -status <serial>' prints the status of the cert with
3728 the given serial number (according to the index file).
3729 'openssl ca -updatedb' updates the expiry status of certificates
3731 [Massimiliano Pala <madwolf@comune.modena.it>]
3733 *) New '-newreq-nodes' command option to CA.pl. This is like
3734 '-newreq', but calls 'openssl req' with the '-nodes' option
3735 so that the resulting key is not encrypted.
3736 [Damien Miller <djm@mindrot.org>]
3738 *) New configuration for the GNU Hurd.
3739 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
3741 *) Initial code to implement OCSP basic response verify. This
3742 is currently incomplete. Currently just finds the signer's
3743 certificate and verifies the signature on the response.
3746 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
3747 value of OPENSSLDIR. This is available via the new '-d' option
3748 to 'openssl version', and is also included in 'openssl version -a'.
3751 *) Allowing defining memory allocation callbacks that will be given
3752 file name and line number information in additional arguments
3753 (a const char* and an int). The basic functionality remains, as
3754 well as the original possibility to just replace malloc(),
3755 realloc() and free() by functions that do not know about these
3756 additional arguments. To register and find out the current
3757 settings for extended allocation functions, the following
3758 functions are provided:
3760 CRYPTO_set_mem_ex_functions
3761 CRYPTO_set_locked_mem_ex_functions
3762 CRYPTO_get_mem_ex_functions
3763 CRYPTO_get_locked_mem_ex_functions
3765 These work the same way as CRYPTO_set_mem_functions and friends.
3766 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
3767 extended allocation function is enabled.
3768 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
3769 a conventional allocation function is enabled.
3770 [Richard Levitte, Bodo Moeller]
3772 *) Finish off removing the remaining LHASH function pointer casts.
3773 There should no longer be any prototype-casting required when using
3774 the LHASH abstraction, and any casts that remain are "bugs". See
3775 the callback types and macros at the head of lhash.h for details
3776 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
3779 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
3780 If /dev/[u]random devices are not available or do not return enough
3781 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
3783 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
3784 /etc/entropy will be queried once each in this sequence, quering stops
3785 when enough entropy was collected without querying more sockets.
3788 *) Change the Unix RAND_poll() variant to be able to poll several
3789 random devices, as specified by DEVRANDOM, until a sufficient amount
3790 of data has been collected. We spend at most 10 ms on each file
3791 (select timeout) and read in non-blocking mode. DEVRANDOM now
3792 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
3793 (previously it was just the string "/dev/urandom"), so on typical
3794 platforms the 10 ms delay will never occur.
3795 Also separate out the Unix variant to its own file, rand_unix.c.
3796 For VMS, there's a currently-empty rand_vms.c.
3799 *) Move OCSP client related routines to ocsp_cl.c. These
3800 provide utility functions which an application needing
3801 to issue a request to an OCSP responder and analyse the
3802 response will typically need: as opposed to those which an
3803 OCSP responder itself would need which will be added later.
3805 OCSP_request_sign() signs an OCSP request with an API similar
3806 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
3807 response. OCSP_response_get1_basic() extracts basic response
3808 from response. OCSP_resp_find_status(): finds and extracts status
3809 information from an OCSP_CERTID structure (which will be created
3810 when the request structure is built). These are built from lower
3811 level functions which work on OCSP_SINGLERESP structures but
3812 wont normally be used unless the application wishes to examine
3813 extensions in the OCSP response for example.
3815 Replace nonce routines with a pair of functions.
3816 OCSP_request_add1_nonce() adds a nonce value and optionally
3817 generates a random value. OCSP_check_nonce() checks the
3818 validity of the nonce in an OCSP response.
3821 *) Change function OCSP_request_add() to OCSP_request_add0_id().
3822 This doesn't copy the supplied OCSP_CERTID and avoids the
3823 need to free up the newly created id. Change return type
3824 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
3825 This can then be used to add extensions to the request.
3826 Deleted OCSP_request_new(), since most of its functionality
3827 is now in OCSP_REQUEST_new() (and the case insensitive name
3828 clash) apart from the ability to set the request name which
3829 will be added elsewhere.
3832 *) Update OCSP API. Remove obsolete extensions argument from
3833 various functions. Extensions are now handled using the new
3834 OCSP extension code. New simple OCSP HTTP function which
3835 can be used to send requests and parse the response.
3838 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
3839 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
3840 uses the special reorder version of SET OF to sort the attributes
3841 and reorder them to match the encoded order. This resolves a long
3842 standing problem: a verify on a PKCS7 structure just after signing
3843 it used to fail because the attribute order did not match the
3844 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
3845 it uses the received order. This is necessary to tolerate some broken
3846 software that does not order SET OF. This is handled by encoding
3847 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
3848 to produce the required SET OF.
3851 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
3852 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
3853 files to get correct declarations of the ASN.1 item variables.
3856 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
3857 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
3858 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
3859 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
3860 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
3861 ASN1_ITEM and no wrapper functions.
3864 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
3865 replace the old function pointer based I/O routines. Change most of
3866 the *_d2i_bio() and *_d2i_fp() functions to use these.
3869 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
3870 lines, recognice more "algorithms" that can be deselected, and make
3871 it complain about algorithm deselection that isn't recognised.
3874 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
3875 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
3876 to use new functions. Add NO_ASN1_OLD which can be set to remove
3877 some old style ASN1 functions: this can be used to determine if old
3878 code will still work when these eventually go away.
3881 *) New extension functions for OCSP structures, these follow the
3882 same conventions as certificates and CRLs.
3885 *) New function X509V3_add1_i2d(). This automatically encodes and
3886 adds an extension. Its behaviour can be customised with various
3887 flags to append, replace or delete. Various wrappers added for
3888 certifcates and CRLs.
3891 *) Fix to avoid calling the underlying ASN1 print routine when
3892 an extension cannot be parsed. Correct a typo in the
3893 OCSP_SERVICELOC extension. Tidy up print OCSP format.
3896 *) Make mkdef.pl parse some of the ASN1 macros and add apropriate
3897 entries for variables.
3900 *) Add functionality to apps/openssl.c for detecting locking
3901 problems: As the program is single-threaded, all we have
3902 to do is register a locking callback using an array for
3903 storing which locks are currently held by the program.
3906 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
3907 SSL_get_ex_data_X509_STORE_idx(), which is used in
3908 ssl_verify_cert_chain() and thus can be called at any time
3909 during TLS/SSL handshakes so that thread-safety is essential.
3910 Unfortunately, the ex_data design is not at all suited
3911 for multi-threaded use, so it probably should be abolished.
3914 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
3915 [Broadcom, tweaked and integrated by Geoff Thorpe]
3917 *) Move common extension printing code to new function
3918 X509V3_print_extensions(). Reorganise OCSP print routines and
3919 implement some needed OCSP ASN1 functions. Add OCSP extensions.
3922 *) New function X509_signature_print() to remove duplication in some
3926 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
3927 set (this was treated exactly the same as SET OF previously). This
3928 is used to reorder the STACK representing the structure to match the
3929 encoding. This will be used to get round a problem where a PKCS7
3930 structure which was signed could not be verified because the STACK
3931 order did not reflect the encoded order.
3934 *) Reimplement the OCSP ASN1 module using the new code.
3937 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
3938 for its ASN1 operations. The old style function pointers still exist
3939 for now but they will eventually go away.
3942 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
3943 completely replaces the old ASN1 functionality with a table driven
3944 encoder and decoder which interprets an ASN1_ITEM structure describing
3945 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
3946 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
3947 has also been converted to the new form.
3950 *) Change BN_mod_exp_recp so that negative moduli are tolerated
3951 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
3952 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
3953 for negative moduli.
3956 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
3957 of not touching the result's sign bit.
3960 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
3964 *) Changed the LHASH code to use prototypes for callbacks, and created
3965 macros to declare and implement thin (optionally static) functions
3966 that provide type-safety and avoid function pointer casting for the
3967 type-specific callbacks.
3970 *) Added Kerberos Cipher Suites to be used with TLS, as written in
3972 [Veers Staats <staatsvr@asc.hpc.mil>,
3973 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
3975 *) Reformat the FAQ so the different questions and answers can be divided
3976 in sections depending on the subject.
3979 *) Have the zlib compression code load ZLIB.DLL dynamically under
3983 *) New function BN_mod_sqrt for computing square roots modulo a prime
3984 (using the probabilistic Tonelli-Shanks algorithm unless
3985 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
3986 be handled deterministically).
3987 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
3989 *) Make BN_mod_inverse faster by explicitly handling small quotients
3990 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
3991 512 bits], about 30% for larger ones [1024 or 2048 bits].)
3994 *) New function BN_kronecker.
3997 *) Fix BN_gcd so that it works on negative inputs; the result is
3998 positive unless both parameters are zero.
3999 Previously something reasonably close to an infinite loop was
4000 possible because numbers could be growing instead of shrinking
4001 in the implementation of Euclid's algorithm.
4004 *) Fix BN_is_word() and BN_is_one() macros to take into account the
4005 sign of the number in question.
4007 Fix BN_is_word(a,w) to work correctly for w == 0.
4009 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
4010 because its test if the absolute value of 'a' equals 'w'.
4011 Note that BN_abs_is_word does *not* handle w == 0 reliably;
4012 it exists mostly for use in the implementations of BN_is_zero(),
4013 BN_is_one(), and BN_is_word().
4016 *) New function BN_swap.
4019 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
4020 the exponentiation functions are more likely to produce reasonable
4021 results on negative inputs.
4024 *) Change BN_mod_mul so that the result is always non-negative.
4025 Previously, it could be negative if one of the factors was negative;
4026 I don't think anyone really wanted that behaviour.
4029 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
4030 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
4031 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
4032 and add new functions:
4041 BN_mod_lshift1_quick
4045 These functions always generate non-negative results.
4047 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
4048 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
4050 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
4051 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
4052 be reduced modulo m.
4053 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
4056 The following entry accidentily appeared in the CHANGES file
4057 distributed with OpenSSL 0.9.7. The modifications described in
4058 it do *not* apply to OpenSSL 0.9.7.
4060 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
4061 was actually never needed) and in BN_mul(). The removal in BN_mul()
4062 required a small change in bn_mul_part_recursive() and the addition
4063 of the functions bn_cmp_part_words(), bn_sub_part_words() and
4064 bn_add_part_words(), which do the same thing as bn_cmp_words(),
4065 bn_sub_words() and bn_add_words() except they take arrays with
4070 *) In 'openssl passwd', verify passwords read from the terminal
4071 unless the '-salt' option is used (which usually means that
4072 verification would just waste user's time since the resulting
4073 hash is going to be compared with some given password hash)
4074 or the new '-noverify' option is used.
4076 This is an incompatible change, but it does not affect
4077 non-interactive use of 'openssl passwd' (passwords on the command
4078 line, '-stdin' option, '-in ...' option) and thus should not
4082 *) Remove all references to RSAref, since there's no more need for it.
4085 *) Make DSO load along a path given through an environment variable
4086 (SHLIB_PATH) with shl_load().
4089 *) Constify the ENGINE code as a result of BIGNUM constification.
4090 Also constify the RSA code and most things related to it. In a
4091 few places, most notable in the depth of the ASN.1 code, ugly
4092 casts back to non-const were required (to be solved at a later
4096 *) Make it so the openssl application has all engines loaded by default.
4099 *) Constify the BIGNUM routines a little more.
4102 *) Add the following functions:
4104 ENGINE_load_cswift()
4106 ENGINE_load_atalla()
4108 ENGINE_load_builtin_engines()
4110 That way, an application can itself choose if external engines that
4111 are built-in in OpenSSL shall ever be used or not. The benefit is
4112 that applications won't have to be linked with libdl or other dso
4113 libraries unless it's really needed.
4115 Changed 'openssl engine' to load all engines on demand.
4116 Changed the engine header files to avoid the duplication of some
4117 declarations (they differed!).
4120 *) 'openssl engine' can now list capabilities.
4123 *) Better error reporting in 'openssl engine'.
4126 *) Never call load_dh_param(NULL) in s_server.
4129 *) Add engine application. It can currently list engines by name and
4130 identity, and test if they are actually available.
4133 *) Improve RPM specification file by forcing symbolic linking and making
4134 sure the installed documentation is also owned by root.root.
4135 [Damien Miller <djm@mindrot.org>]
4137 *) Give the OpenSSL applications more possibilities to make use of
4138 keys (public as well as private) handled by engines.
4141 *) Add OCSP code that comes from CertCo.
4144 *) Add VMS support for the Rijndael code.
4147 *) Added untested support for Nuron crypto accelerator.
4150 *) Add support for external cryptographic devices. This code was
4151 previously distributed separately as the "engine" branch.
4152 [Geoff Thorpe, Richard Levitte]
4154 *) Rework the filename-translation in the DSO code. It is now possible to
4155 have far greater control over how a "name" is turned into a filename
4156 depending on the operating environment and any oddities about the
4157 different shared library filenames on each system.
4160 *) Support threads on FreeBSD-elf in Configure.
4163 *) Fix for SHA1 assembly problem with MASM: it produces
4164 warnings about corrupt line number information when assembling
4165 with debugging information. This is caused by the overlapping
4167 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
4170 NCONF_get_number() has no error checking at all. As a replacement,
4171 NCONF_get_number_e() is defined (_e for "error checking") and is
4172 promoted strongly. The old NCONF_get_number is kept around for
4173 binary backward compatibility.
4174 Make it possible for methods to load from something other than a BIO,
4175 by providing a function pointer that is given a name instead of a BIO.
4176 For example, this could be used to load configuration data from an
4180 *) Fix for non blocking accept BIOs. Added new I/O special reason
4181 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
4182 with non blocking I/O was not possible because no retry code was
4183 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
4187 *) Added the beginnings of Rijndael support.
4190 *) Fix for bug in DirectoryString mask setting. Add support for
4191 X509_NAME_print_ex() in 'req' and X509_print_ex() function
4192 to allow certificate printing to more controllable, additional
4193 'certopt' option to 'x509' to allow new printing options to be
4197 *) Clean old EAY MD5 hack from e_os.h.
4200 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
4202 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
4203 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
4204 [Joe Orton, Steve Henson]
4206 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
4208 *) Fix additional bug revealed by the NISCC test suite:
4210 Stop bug triggering large recursion when presented with
4211 certain ASN.1 tags (CVE-2003-0851)
4214 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
4216 *) Fix various bugs revealed by running the NISCC test suite:
4218 Stop out of bounds reads in the ASN1 code when presented with
4219 invalid tags (CVE-2003-0543 and CVE-2003-0544).
4221 If verify callback ignores invalid public key errors don't try to check
4222 certificate signature with the NULL public key.
4226 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
4227 if the server requested one: as stated in TLS 1.0 and SSL 3.0
4231 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
4232 extra data after the compression methods not only for TLS 1.0
4233 but also for SSL 3.0 (as required by the specification).
4234 [Bodo Moeller; problem pointed out by Matthias Loepfe]
4236 *) Change X509_certificate_type() to mark the key as exported/exportable
4237 when it's 512 *bits* long, not 512 bytes.
4240 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
4242 *) Countermeasure against the Klima-Pokorny-Rosa extension of
4243 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
4244 a protocol version number mismatch like a decryption error
4245 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
4248 *) Turn on RSA blinding by default in the default implementation
4249 to avoid a timing attack. Applications that don't want it can call
4250 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
4251 They would be ill-advised to do so in most cases.
4252 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
4254 *) Change RSA blinding code so that it works when the PRNG is not
4255 seeded (in this case, the secret RSA exponent is abused as
4256 an unpredictable seed -- if it is not unpredictable, there
4257 is no point in blinding anyway). Make RSA blinding thread-safe
4258 by remembering the creator's thread ID in rsa->blinding and
4259 having all other threads use local one-time blinding factors
4260 (this requires more computation than sharing rsa->blinding, but
4261 avoids excessive locking; and if an RSA object is not shared
4262 between threads, blinding will still be very fast).
4265 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
4267 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
4268 via timing by performing a MAC computation even if incorrrect
4269 block cipher padding has been found. This is a countermeasure
4270 against active attacks where the attacker has to distinguish
4271 between bad padding and a MAC verification error. (CVE-2003-0078)
4273 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
4274 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
4275 Martin Vuagnoux (EPFL, Ilion)]
4277 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
4279 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
4280 memory from it's contents. This is done with a counter that will
4281 place alternating values in each byte. This can be used to solve
4282 two issues: 1) the removal of calls to memset() by highly optimizing
4283 compilers, and 2) cleansing with other values than 0, since those can
4284 be read through on certain media, for example a swap space on disk.
4287 *) Bugfix: client side session caching did not work with external caching,
4288 because the session->cipher setting was not restored when reloading
4289 from the external cache. This problem was masked, when
4290 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
4291 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
4294 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
4295 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
4296 [Zeev Lieber <zeev-l@yahoo.com>]
4298 *) Undo an undocumented change introduced in 0.9.6e which caused
4299 repeated calls to OpenSSL_add_all_ciphers() and
4300 OpenSSL_add_all_digests() to be ignored, even after calling
4304 *) Change the default configuration reader to deal with last line not
4305 being properly terminated.
4308 *) Change X509_NAME_cmp() so it applies the special rules on handling
4309 DN values that are of type PrintableString, as well as RDNs of type
4310 emailAddress where the value has the type ia5String.
4311 [stefank@valicert.com via Richard Levitte]
4313 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
4314 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
4315 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
4316 the bitwise-OR of the two for use by the majority of applications
4317 wanting this behaviour, and update the docs. The documented
4318 behaviour and actual behaviour were inconsistent and had been
4319 changing anyway, so this is more a bug-fix than a behavioural
4321 [Geoff Thorpe, diagnosed by Nadav Har'El]
4323 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
4324 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
4327 *) Fix initialization code race conditions in
4328 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
4329 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
4330 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
4331 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
4332 ssl2_get_cipher_by_char(),
4333 ssl3_get_cipher_by_char().
4334 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
4336 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
4337 the cached sessions are flushed, as the remove_cb() might use ex_data
4338 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
4339 (see [openssl.org #212]).
4340 [Geoff Thorpe, Lutz Jaenicke]
4342 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
4343 length, instead of the encoding length to d2i_ASN1_OBJECT.
4346 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
4348 *) [In 0.9.6g-engine release:]
4349 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
4350 [Lynn Gazis <lgazis@rainbow.com>]
4352 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
4354 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
4355 and get fix the header length calculation.
4356 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
4357 Alon Kantor <alonk@checkpoint.com> (and others),
4360 *) Use proper error handling instead of 'assertions' in buffer
4361 overflow checks added in 0.9.6e. This prevents DoS (the
4362 assertions could call abort()).
4363 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
4365 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
4367 *) Add various sanity checks to asn1_get_length() to reject
4368 the ASN1 length bytes if they exceed sizeof(long), will appear
4369 negative or the content length exceeds the length of the
4371 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4373 *) Fix cipher selection routines: ciphers without encryption had no flags
4374 for the cipher strength set and where therefore not handled correctly
4375 by the selection routines (PR #130).
4378 *) Fix EVP_dsa_sha macro.
4382 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
4383 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
4384 that was added in OpenSSL 0.9.6d.
4386 As the countermeasure turned out to be incompatible with some
4387 broken SSL implementations, the new option is part of SSL_OP_ALL.
4388 SSL_OP_ALL is usually employed when compatibility with weird SSL
4389 implementations is desired (e.g. '-bugs' option to 's_client' and
4390 's_server'), so the new option is automatically set in many
4394 *) Changes in security patch:
4396 Changes marked "(CHATS)" were sponsored by the Defense Advanced
4397 Research Projects Agency (DARPA) and Air Force Research Laboratory,
4398 Air Force Materiel Command, USAF, under agreement number
4401 *) Add various sanity checks to asn1_get_length() to reject
4402 the ASN1 length bytes if they exceed sizeof(long), will appear
4403 negative or the content length exceeds the length of the
4404 supplied buffer. (CVE-2002-0659)
4405 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
4407 *) Assertions for various potential buffer overflows, not known to
4409 [Ben Laurie (CHATS)]
4411 *) Various temporary buffers to hold ASCII versions of integers were
4412 too small for 64 bit platforms. (CVE-2002-0655)
4413 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
4415 *) Remote buffer overflow in SSL3 protocol - an attacker could
4416 supply an oversized session ID to a client. (CVE-2002-0656)
4417 [Ben Laurie (CHATS)]
4419 *) Remote buffer overflow in SSL2 protocol - an attacker could
4420 supply an oversized client master key. (CVE-2002-0656)
4421 [Ben Laurie (CHATS)]
4423 Changes between 0.9.6c and 0.9.6d [9 May 2002]
4425 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
4426 encoded as NULL) with id-dsa-with-sha1.
4427 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
4429 *) Check various X509_...() return values in apps/req.c.
4430 [Nils Larsch <nla@trustcenter.de>]
4432 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
4433 an end-of-file condition would erronously be flagged, when the CRLF
4434 was just at the end of a processed block. The bug was discovered when
4435 processing data through a buffering memory BIO handing the data to a
4436 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
4437 <ptsekov@syntrex.com> and Nedelcho Stanev.
4440 *) Implement a countermeasure against a vulnerability recently found
4441 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
4442 before application data chunks to avoid the use of known IVs
4443 with data potentially chosen by the attacker.
4446 *) Fix length checks in ssl3_get_client_hello().
4449 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
4450 to prevent ssl3_read_internal() from incorrectly assuming that
4451 ssl3_read_bytes() found application data while handshake
4452 processing was enabled when in fact s->s3->in_read_app_data was
4453 merely automatically cleared during the initial handshake.
4454 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
4456 *) Fix object definitions for Private and Enterprise: they were not
4457 recognized in their shortname (=lowercase) representation. Extend
4458 obj_dat.pl to issue an error when using undefined keywords instead
4459 of silently ignoring the problem (Svenning Sorensen
4460 <sss@sss.dnsalias.net>).
4463 *) Fix DH_generate_parameters() so that it works for 'non-standard'
4464 generators, i.e. generators other than 2 and 5. (Previously, the
4465 code did not properly initialise the 'add' and 'rem' values to
4466 BN_generate_prime().)
4468 In the new general case, we do not insist that 'generator' is
4469 actually a primitive root: This requirement is rather pointless;
4470 a generator of the order-q subgroup is just as good, if not
4474 *) Map new X509 verification errors to alerts. Discovered and submitted by
4475 Tom Wu <tom@arcot.com>.
4478 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
4479 returning non-zero before the data has been completely received
4480 when using non-blocking I/O.
4481 [Bodo Moeller; problem pointed out by John Hughes]
4483 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
4484 [Ben Laurie, Lutz Jaenicke]
4486 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
4487 Yoram Zahavi <YoramZ@gilian.com>).
4490 *) Add information about CygWin 1.3 and on, and preserve proper
4491 configuration for the versions before that.
4492 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
4494 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
4495 check whether we deal with a copy of a session and do not delete from
4496 the cache in this case. Problem reported by "Izhar Shoshani Levi"
4497 <izhar@checkpoint.com>.
4500 *) Do not store session data into the internal session cache, if it
4501 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
4502 flag is set). Proposed by Aslam <aslam@funk.com>.
4505 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
4509 *) [In 0.9.6d-engine release:]
4510 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
4511 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
4513 *) Add the configuration target linux-s390x.
4514 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
4516 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
4517 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
4518 variable as an indication that a ClientHello message has been
4519 received. As the flag value will be lost between multiple
4520 invocations of ssl3_accept when using non-blocking I/O, the
4521 function may not be aware that a handshake has actually taken
4522 place, thus preventing a new session from being added to the
4525 To avoid this problem, we now set s->new_session to 2 instead of
4526 using a local variable.
4527 [Lutz Jaenicke, Bodo Moeller]
4529 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
4530 if the SSL_R_LENGTH_MISMATCH error is detected.
4531 [Geoff Thorpe, Bodo Moeller]
4533 *) New 'shared_ldflag' column in Configure platform table.
4536 *) Fix EVP_CIPHER_mode macro.
4537 ["Dan S. Camper" <dan@bti.net>]
4539 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
4540 type, we must throw them away by setting rr->length to 0.
4541 [D P Chang <dpc@qualys.com>]
4543 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
4545 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
4546 <Dominikus.Scherkl@biodata.com>. (The previous implementation
4547 worked incorrectly for those cases where range = 10..._2 and
4548 3*range is two bits longer than range.)
4551 *) Only add signing time to PKCS7 structures if it is not already
4555 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
4556 OBJ_ld_ce should be OBJ_id_ce.
4557 Also some ip-pda OIDs in crypto/objects/objects.txt were
4558 incorrect (cf. RFC 3039).
4559 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
4561 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
4562 returns early because it has nothing to do.
4563 [Andy Schneider <andy.schneider@bjss.co.uk>]
4565 *) [In 0.9.6c-engine release:]
4566 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
4567 [Andy Schneider <andy.schneider@bjss.co.uk>]
4569 *) [In 0.9.6c-engine release:]
4570 Add support for Cryptographic Appliance's keyserver technology.
4571 (Use engine 'keyclient')
4572 [Cryptographic Appliances and Geoff Thorpe]
4574 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
4575 is called via tools/c89.sh because arguments have to be
4576 rearranged (all '-L' options must appear before the first object
4578 [Richard Shapiro <rshapiro@abinitio.com>]
4580 *) [In 0.9.6c-engine release:]
4581 Add support for Broadcom crypto accelerator cards, backported
4583 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
4585 *) [In 0.9.6c-engine release:]
4586 Add support for SureWare crypto accelerator cards from
4587 Baltimore Technologies. (Use engine 'sureware')
4588 [Baltimore Technologies and Mark Cox]
4590 *) [In 0.9.6c-engine release:]
4591 Add support for crypto accelerator cards from Accelerated
4592 Encryption Processing, www.aep.ie. (Use engine 'aep')
4593 [AEP Inc. and Mark Cox]
4595 *) Add a configuration entry for gcc on UnixWare.
4596 [Gary Benson <gbenson@redhat.com>]
4598 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
4599 messages are stored in a single piece (fixed-length part and
4600 variable-length part combined) and fix various bugs found on the way.
4603 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
4604 instead. BIO_gethostbyname() does not know what timeouts are
4605 appropriate, so entries would stay in cache even when they have
4607 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
4609 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
4610 faced with a pathologically small ClientHello fragment that does
4611 not contain client_version: Instead of aborting with an error,
4612 simply choose the highest available protocol version (i.e.,
4613 TLS 1.0 unless it is disabled). In practice, ClientHello
4614 messages are never sent like this, but this change gives us
4615 strictly correct behaviour at least for TLS.
4618 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
4619 never resets s->method to s->ctx->method when called from within
4620 one of the SSL handshake functions.
4621 [Bodo Moeller; problem pointed out by Niko Baric]
4623 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
4624 (sent using the client's version number) if client_version is
4625 smaller than the protocol version in use. Also change
4626 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
4627 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
4628 the client will at least see that alert.
4631 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
4635 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
4636 client receives HelloRequest while in a handshake.
4637 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
4639 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
4640 should end in 'break', not 'goto end' which circuments various
4641 cleanups done in state SSL_ST_OK. But session related stuff
4642 must be disabled for SSL_ST_OK in the case that we just sent a
4645 Also avoid some overhead by not calling ssl_init_wbio_buffer()
4646 before just sending a HelloRequest.
4647 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
4649 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
4650 reveal whether illegal block cipher padding was found or a MAC
4651 verification error occured. (Neither SSLerr() codes nor alerts
4652 are directly visible to potential attackers, but the information
4653 may leak via logfiles.)
4655 Similar changes are not required for the SSL 2.0 implementation
4656 because the number of padding bytes is sent in clear for SSL 2.0,
4657 and the extra bytes are just ignored. However ssl/s2_pkt.c
4658 failed to verify that the purported number of padding bytes is in
4662 *) Add OpenUNIX-8 support including shared libraries
4663 (Boyd Lynn Gerber <gerberb@zenez.com>).
4666 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
4667 'wristwatch attack' using huge encoding parameters (cf.
4668 James H. Manger's CRYPTO 2001 paper). Note that the
4669 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
4670 encoding parameters and hence was not vulnerable.
4673 *) BN_sqr() bug fix.
4674 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
4676 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
4677 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
4678 followed by modular reduction.
4679 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
4681 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
4682 equivalent based on BN_pseudo_rand() instead of BN_rand().
4685 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
4686 This function was broken, as the check for a new client hello message
4687 to handle SGC did not allow these large messages.
4688 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
4691 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
4694 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
4695 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
4698 *) Rework the configuration and shared library support for Tru64 Unix.
4699 The configuration part makes use of modern compiler features and
4700 still retains old compiler behavior for those that run older versions
4701 of the OS. The shared library support part includes a variant that
4702 uses the RPATH feature, and is available through the special
4703 configuration target "alpha-cc-rpath", which will never be selected
4705 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
4707 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
4708 with the same message size as in ssl3_get_certificate_request().
4709 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
4710 messages might inadvertently be reject as too long.
4711 [Petr Lampa <lampa@fee.vutbr.cz>]
4713 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
4716 *) Modified SSL library such that the verify_callback that has been set
4717 specificly for an SSL object with SSL_set_verify() is actually being
4718 used. Before the change, a verify_callback set with this function was
4719 ignored and the verify_callback() set in the SSL_CTX at the time of
4720 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
4721 to allow the necessary settings.
4724 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
4725 explicitly to NULL, as at least on Solaris 8 this seems not always to be
4726 done automatically (in contradiction to the requirements of the C
4727 standard). This made problems when used from OpenSSH.
4730 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
4731 dh->length and always used
4733 BN_rand_range(priv_key, dh->p).
4735 BN_rand_range() is not necessary for Diffie-Hellman, and this
4736 specific range makes Diffie-Hellman unnecessarily inefficient if
4737 dh->length (recommended exponent length) is much smaller than the
4738 length of dh->p. We could use BN_rand_range() if the order of
4739 the subgroup was stored in the DH structure, but we only have
4744 BN_rand(priv_key, l, ...)
4746 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
4752 RSA_eay_public_encrypt
4753 RSA_eay_private_decrypt
4754 RSA_eay_private_encrypt (signing)
4755 RSA_eay_public_decrypt (signature verification)
4757 (default implementations for RSA_public_encrypt,
4758 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
4759 always reject numbers >= n.
4762 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
4763 to synchronize access to 'locking_thread'. This is necessary on
4764 systems where access to 'locking_thread' (an 'unsigned long'
4765 variable) is not atomic.
4768 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
4769 *before* setting the 'crypto_lock_rand' flag. The previous code had
4770 a race condition if 0 is a valid thread ID.
4771 [Travis Vitek <vitek@roguewave.com>]
4773 *) Add support for shared libraries under Irix.
4774 [Albert Chin-A-Young <china@thewrittenword.com>]
4776 *) Add configuration option to build on Linux on both big-endian and
4778 [Ralf Baechle <ralf@uni-koblenz.de>]
4780 *) Add the possibility to create shared libraries on HP-UX.
4783 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
4785 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
4786 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
4787 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
4788 PRNG state recovery was possible based on the output of
4789 one PRNG request appropriately sized to gain knowledge on
4790 'md' followed by enough consecutive 1-byte PRNG requests
4791 to traverse all of 'state'.
4793 1. When updating 'md_local' (the current thread's copy of 'md')
4794 during PRNG output generation, hash all of the previous
4795 'md_local' value, not just the half used for PRNG output.
4797 2. Make the number of bytes from 'state' included into the hash
4798 independent from the number of PRNG bytes requested.
4800 The first measure alone would be sufficient to avoid
4801 Markku-Juhani's attack. (Actually it had never occurred
4802 to me that the half of 'md_local' used for chaining was the
4803 half from which PRNG output bytes were taken -- I had always
4804 assumed that the secret half would be used.) The second
4805 measure makes sure that additional data from 'state' is never
4806 mixed into 'md_local' in small portions; this heuristically
4807 further strengthens the PRNG.
4810 *) Fix crypto/bn/asm/mips3.s.
4813 *) When only the key is given to "enc", the IV is undefined. Print out
4814 an error message in this case.
4817 *) Handle special case when X509_NAME is empty in X509 printing routines.
4820 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
4821 positive and less than q.
4824 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
4825 used: it isn't thread safe and the add_lock_callback should handle
4827 [Paul Rose <Paul.Rose@bridge.com>]
4829 *) Verify that incoming data obeys the block size in
4830 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
4834 [Ulf Möller, Bodo Möller]
4836 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
4837 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
4838 when fixing the server behaviour for backwards-compatible 'client
4839 hello' messages. (Note that the attack is impractical against
4840 SSL 3.0 and TLS 1.0 anyway because length and version checking
4841 means that the probability of guessing a valid ciphertext is
4842 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
4845 Before 0.9.5, the countermeasure (hide the error by generating a
4846 random 'decryption result') did not work properly because
4847 ERR_clear_error() was missing, meaning that SSL_get_error() would
4848 detect the supposedly ignored error.
4850 Both problems are now fixed.
4853 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
4854 (previously it was 1024).
4857 *) Fix for compatibility mode trust settings: ignore trust settings
4858 unless some valid trust or reject settings are present.
4861 *) Fix for blowfish EVP: its a variable length cipher.
4864 *) Fix various bugs related to DSA S/MIME verification. Handle missing
4865 parameters in DSA public key structures and return an error in the
4866 DSA routines if parameters are absent.
4869 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
4870 in the current directory if neither $RANDFILE nor $HOME was set.
4871 RAND_file_name() in 0.9.6a returned NULL in this case. This has
4872 caused some confusion to Windows users who haven't defined $HOME.
4873 Thus RAND_file_name() is changed again: e_os.h can define a
4874 DEFAULT_HOME, which will be used if $HOME is not set.
4875 For Windows, we use "C:"; on other platforms, we still require
4876 environment variables.
4878 *) Move 'if (!initialized) RAND_poll()' into regions protected by
4879 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
4880 having multiple threads call RAND_poll() concurrently.
4883 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
4884 combination of a flag and a thread ID variable.
4885 Otherwise while one thread is in ssleay_rand_bytes (which sets the
4886 flag), *other* threads can enter ssleay_add_bytes without obeying
4887 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
4888 that they do not hold after the first thread unsets add_do_not_lock).
4891 *) Change bctest again: '-x' expressions are not available in all
4895 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
4897 *) Fix a couple of memory leaks in PKCS7_dataDecode()
4898 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
4900 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
4901 the default extension for executables, if any. Also, make the perl
4902 scripts that use symlink() to test if it really exists and use "cp"
4903 if it doesn't. All this made OpenSSL compilable and installable in
4907 *) Fix for asn1_GetSequence() for indefinite length constructed data.
4908 If SEQUENCE is length is indefinite just set c->slen to the total
4909 amount of data available.
4910 [Steve Henson, reported by shige@FreeBSD.org]
4911 [This change does not apply to 0.9.7.]
4913 *) Change bctest to avoid here-documents inside command substitution
4914 (workaround for FreeBSD /bin/sh bug).
4915 For compatibility with Ultrix, avoid shell functions (introduced
4916 in the bctest version that searches along $PATH).
4919 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
4920 with des_encrypt() defined on some operating systems, like Solaris
4924 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
4925 On the Importance of Eliminating Errors in Cryptographic
4926 Computations, J. Cryptology 14 (2001) 2, 101-119,
4927 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
4930 *) MIPS assembler BIGNUM division bug fix.
4933 *) Disabled incorrect Alpha assembler code.
4936 *) Fix PKCS#7 decode routines so they correctly update the length
4937 after reading an EOC for the EXPLICIT tag.
4939 [This change does not apply to 0.9.7.]
4941 *) Fix bug in PKCS#12 key generation routines. This was triggered
4942 if a 3DES key was generated with a 0 initial byte. Include
4943 PKCS12_BROKEN_KEYGEN compilation option to retain the old
4944 (but broken) behaviour.
4947 *) Enhance bctest to search for a working bc along $PATH and print
4949 [Tim Rice <tim@multitalents.net> via Richard Levitte]
4951 *) Fix memory leaks in err.c: free err_data string if necessary;
4952 don't write to the wrong index in ERR_set_error_data.
4955 *) Implement ssl23_peek (analogous to ssl23_read), which previously
4959 *) Replace rdtsc with _emit statements for VC++ version 5.
4960 [Jeremy Cooper <jeremy@baymoo.org>]
4962 *) Make it possible to reuse SSLv2 sessions.
4965 *) In copy_email() check for >= 0 as a return value for
4966 X509_NAME_get_index_by_NID() since 0 is a valid index.
4967 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
4969 *) Avoid coredump with unsupported or invalid public keys by checking if
4970 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
4971 PKCS7_verify() fails with non detached data.
4974 *) Don't use getenv in library functions when run as setuid/setgid.
4975 New function OPENSSL_issetugid().
4978 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
4979 due to incorrect handling of multi-threading:
4981 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
4983 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
4985 3. Count how many times MemCheck_off() has been called so that
4986 nested use can be treated correctly. This also avoids
4987 inband-signalling in the previous code (which relied on the
4988 assumption that thread ID 0 is impossible).
4991 *) Add "-rand" option also to s_client and s_server.
4994 *) Fix CPU detection on Irix 6.x.
4995 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
4996 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
4998 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
5001 [This change does not apply to 0.9.7.]
5003 *) Use the cached encoding of an X509_NAME structure rather than
5004 copying it. This is apparently the reason for the libsafe "errors"
5005 but the code is actually correct.
5008 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
5009 Bleichenbacher's DSA attack.
5010 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
5011 to be set and top=0 forces the highest bit to be set; top=-1 is new
5012 and leaves the highest bit random.
5013 [Ulf Moeller, Bodo Moeller]
5015 *) In the NCONF_...-based implementations for CONF_... queries
5016 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
5017 a temporary CONF structure with the data component set to NULL
5018 (which gives segmentation faults in lh_retrieve).
5019 Instead, use NULL for the CONF pointer in CONF_get_string and
5020 CONF_get_number (which may use environment variables) and directly
5021 return NULL from CONF_get_section.
5024 *) Fix potential buffer overrun for EBCDIC.
5027 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
5028 keyUsage if basicConstraints absent for a CA.
5031 *) Make SMIME_write_PKCS7() write mail header values with a format that
5032 is more generally accepted (no spaces before the semicolon), since
5033 some programs can't parse those values properly otherwise. Also make
5034 sure BIO's that break lines after each write do not create invalid
5038 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
5039 macros previously used would not encode an empty SEQUENCE OF
5040 and break the signature.
5042 [This change does not apply to 0.9.7.]
5044 *) Zero the premaster secret after deriving the master secret in
5048 *) Add some EVP_add_digest_alias registrations (as found in
5049 OpenSSL_add_all_digests()) to SSL_library_init()
5050 aka OpenSSL_add_ssl_algorithms(). This provides improved
5051 compatibility with peers using X.509 certificates
5052 with unconventional AlgorithmIdentifier OIDs.
5055 *) Fix for Irix with NO_ASM.
5056 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
5058 *) ./config script fixes.
5059 [Ulf Moeller, Richard Levitte]
5061 *) Fix 'openssl passwd -1'.
5064 *) Change PKCS12_key_gen_asc() so it can cope with non null
5065 terminated strings whose length is passed in the passlen
5066 parameter, for example from PEM callbacks. This was done
5067 by adding an extra length parameter to asc2uni().
5068 [Steve Henson, reported by <oddissey@samsung.co.kr>]
5070 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
5071 call failed, free the DSA structure.
5074 *) Fix to uni2asc() to cope with zero length Unicode strings.
5075 These are present in some PKCS#12 files.
5078 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
5079 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
5080 when writing a 32767 byte record.
5081 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
5083 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
5084 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
5086 (RSA objects have a reference count access to which is protected
5087 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
5088 so they are meant to be shared between threads.)
5089 [Bodo Moeller, Geoff Thorpe; original patch submitted by
5090 "Reddie, Steven" <Steven.Reddie@ca.com>]
5092 *) Fix a deadlock in CRYPTO_mem_leaks().
5095 *) Use better test patterns in bntest.
5098 *) rand_win.c fix for Borland C.
5101 *) BN_rshift bugfix for n == 0.
5104 *) Add a 'bctest' script that checks for some known 'bc' bugs
5105 so that 'make test' does not abort just because 'bc' is broken.
5108 *) Store verify_result within SSL_SESSION also for client side to
5109 avoid potential security hole. (Re-used sessions on the client side
5110 always resulted in verify_result==X509_V_OK, not using the original
5111 result of the server certificate verification.)
5114 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
5115 SSL3_RT_APPLICATION_DATA, return 0.
5116 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
5120 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
5121 releases, have been re-implemented by renaming the previous
5122 implementations of ssl2_read and ssl3_read to ssl2_read_internal
5123 and ssl3_read_internal, respectively, and adding 'peek' parameters
5124 to them. The new ssl[23]_{read,peek} functions are calls to
5125 ssl[23]_read_internal with the 'peek' flag set appropriately.
5126 A 'peek' parameter has also been added to ssl3_read_bytes, which
5127 does the actual work for ssl3_read_internal.
5130 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
5131 the method-specific "init()" handler. Also clean up ex_data after
5132 calling the method-specific "finish()" handler. Previously, this was
5133 happening the other way round.
5136 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
5137 The previous value, 12, was not always sufficient for BN_mod_exp().
5140 *) Make sure that shared libraries get the internal name engine with
5141 the full version number and not just 0. This should mark the
5142 shared libraries as not backward compatible. Of course, this should
5143 be changed again when we can guarantee backward binary compatibility.
5146 *) Fix typo in get_cert_by_subject() in by_dir.c
5147 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
5149 *) Rework the system to generate shared libraries:
5151 - Make note of the expected extension for the shared libraries and
5152 if there is a need for symbolic links from for example libcrypto.so.0
5153 to libcrypto.so.0.9.7. There is extended info in Configure for
5156 - Make as few rebuilds of the shared libraries as possible.
5158 - Still avoid linking the OpenSSL programs with the shared libraries.
5160 - When installing, install the shared libraries separately from the
5164 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
5166 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
5167 and not in SSL_clear because the latter is also used by the
5168 accept/connect functions; previously, the settings made by
5169 SSL_set_read_ahead would be lost during the handshake.
5170 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
5172 *) Correct util/mkdef.pl to be selective about disabled algorithms.
5173 Previously, it would create entries for disableed algorithms no
5177 *) Added several new manual pages for SSL_* function.
5180 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
5182 *) In ssl23_get_client_hello, generate an error message when faced
5183 with an initial SSL 3.0/TLS record that is too small to contain the
5184 first two bytes of the ClientHello message, i.e. client_version.
5185 (Note that this is a pathologic case that probably has never happened
5186 in real life.) The previous approach was to use the version number
5187 from the record header as a substitute; but our protocol choice
5188 should not depend on that one because it is not authenticated
5189 by the Finished messages.
5192 *) More robust randomness gathering functions for Windows.
5193 [Jeffrey Altman <jaltman@columbia.edu>]
5195 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
5196 not set then we don't setup the error code for issuer check errors
5197 to avoid possibly overwriting other errors which the callback does
5198 handle. If an application does set the flag then we assume it knows
5199 what it is doing and can handle the new informational codes
5203 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
5204 a general "ANY" type, as such it should be able to decode anything
5205 including tagged types. However it didn't check the class so it would
5206 wrongly interpret tagged types in the same way as their universal
5207 counterpart and unknown types were just rejected. Changed so that the
5208 tagged and unknown types are handled in the same way as a SEQUENCE:
5209 that is the encoding is stored intact. There is also a new type
5210 "V_ASN1_OTHER" which is used when the class is not universal, in this
5211 case we have no idea what the actual type is so we just lump them all
5215 *) On VMS, stdout may very well lead to a file that is written to
5216 in a record-oriented fashion. That means that every write() will
5217 write a separate record, which will be read separately by the
5218 programs trying to read from it. This can be very confusing.
5220 The solution is to put a BIO filter in the way that will buffer
5221 text until a linefeed is reached, and then write everything a
5222 line at a time, so every record written will be an actual line,
5223 not chunks of lines and not (usually doesn't happen, but I've
5224 seen it once) several lines in one record. BIO_f_linebuffer() is
5227 Currently, it's a VMS-only method, because that's where it has
5228 been tested well enough.
5231 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
5232 it can return incorrect results.
5233 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
5234 but it was in 0.9.6-beta[12].)
5237 *) Disable the check for content being present when verifying detached
5238 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
5239 include zero length content when signing messages.
5242 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
5243 BIO_ctrl (for BIO pairs).
5246 *) Add DSO method for VMS.
5249 *) Bug fix: Montgomery multiplication could produce results with the
5253 *) Add RPM specification openssl.spec and modify it to build three
5254 packages. The default package contains applications, application
5255 documentation and run-time libraries. The devel package contains
5256 include files, static libraries and function documentation. The
5257 doc package contains the contents of the doc directory. The original
5258 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
5261 *) Add a large number of documentation files for many SSL routines.
5262 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
5264 *) Add a configuration entry for Sony News 4.
5265 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
5267 *) Don't set the two most significant bits to one when generating a
5268 random number < q in the DSA library.
5271 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
5272 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
5273 the underlying transport is blocking) if a handshake took place.
5274 (The default behaviour is needed by applications such as s_client
5275 and s_server that use select() to determine when to use SSL_read;
5276 but for applications that know in advance when to expect data, it
5277 just makes things more complicated.)
5280 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
5284 *) Add a few more EBCDIC conditionals that make `req' and `x509'
5285 work better on such systems.
5286 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
5288 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
5289 Update PKCS12_parse() so it copies the friendlyName and the
5290 keyid to the certificates aux info.
5293 *) Fix bug in PKCS7_verify() which caused an infinite loop
5294 if there was more than one signature.
5295 [Sven Uszpelkat <su@celocom.de>]
5297 *) Major change in util/mkdef.pl to include extra information
5298 about each symbol, as well as presentig variables as well
5299 as functions. This change means that there's n more need
5300 to rebuild the .num files when some algorithms are excluded.
5303 *) Allow the verify time to be set by an application,
5304 rather than always using the current time.
5307 *) Phase 2 verify code reorganisation. The certificate
5308 verify code now looks up an issuer certificate by a
5309 number of criteria: subject name, authority key id
5310 and key usage. It also verifies self signed certificates
5311 by the same criteria. The main comparison function is
5312 X509_check_issued() which performs these checks.
5314 Lot of changes were necessary in order to support this
5315 without completely rewriting the lookup code.
5317 Authority and subject key identifier are now cached.
5319 The LHASH 'certs' is X509_STORE has now been replaced
5320 by a STACK_OF(X509_OBJECT). This is mainly because an
5321 LHASH can't store or retrieve multiple objects with
5322 the same hash value.
5324 As a result various functions (which were all internal
5325 use only) have changed to handle the new X509_STORE
5326 structure. This will break anything that messed round
5327 with X509_STORE internally.
5329 The functions X509_STORE_add_cert() now checks for an
5330 exact match, rather than just subject name.
5332 The X509_STORE API doesn't directly support the retrieval
5333 of multiple certificates matching a given criteria, however
5334 this can be worked round by performing a lookup first
5335 (which will fill the cache with candidate certificates)
5336 and then examining the cache for matches. This is probably
5337 the best we can do without throwing out X509_LOOKUP
5338 entirely (maybe later...).
5340 The X509_VERIFY_CTX structure has been enhanced considerably.
5342 All certificate lookup operations now go via a get_issuer()
5343 callback. Although this currently uses an X509_STORE it
5344 can be replaced by custom lookups. This is a simple way
5345 to bypass the X509_STORE hackery necessary to make this
5346 work and makes it possible to use more efficient techniques
5347 in future. A very simple version which uses a simple
5348 STACK for its trusted certificate store is also provided
5349 using X509_STORE_CTX_trusted_stack().
5351 The verify_cb() and verify() callbacks now have equivalents
5352 in the X509_STORE_CTX structure.
5354 X509_STORE_CTX also has a 'flags' field which can be used
5355 to customise the verify behaviour.
5358 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
5359 excludes S/MIME capabilities.
5362 *) When a certificate request is read in keep a copy of the
5363 original encoding of the signed data and use it when outputing
5364 again. Signatures then use the original encoding rather than
5365 a decoded, encoded version which may cause problems if the
5366 request is improperly encoded.
5369 *) For consistency with other BIO_puts implementations, call
5370 buffer_write(b, ...) directly in buffer_puts instead of calling
5373 In BIO_puts, increment b->num_write as in BIO_write.
5374 [Peter.Sylvester@EdelWeb.fr]
5376 *) Fix BN_mul_word for the case where the word is 0. (We have to use
5377 BN_zero, we may not return a BIGNUM with an array consisting of
5381 *) Avoid calling abort() from within the library when problems are
5382 detected, except if preprocessor symbols have been defined
5383 (such as REF_CHECK, BN_DEBUG etc.).
5386 *) New openssl application 'rsautl'. This utility can be
5387 used for low level RSA operations. DER public key
5388 BIO/fp routines also added.
5391 *) New Configure entry and patches for compiling on QNX 4.
5392 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
5394 *) A demo state-machine implementation was sponsored by
5395 Nuron (http://www.nuron.com/) and is now available in
5396 demos/state_machine.
5399 *) New options added to the 'dgst' utility for signature
5400 generation and verification.
5403 *) Unrecognized PKCS#7 content types are now handled via a
5404 catch all ASN1_TYPE structure. This allows unsupported
5405 types to be stored as a "blob" and an application can
5406 encode and decode it manually.
5409 *) Fix various signed/unsigned issues to make a_strex.c
5411 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
5413 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
5414 length if passed a buffer. ASN1_INTEGER_to_BN failed
5415 if passed a NULL BN and its argument was negative.
5416 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
5418 *) Modification to PKCS#7 encoding routines to output definite
5419 length encoding. Since currently the whole structures are in
5420 memory there's not real point in using indefinite length
5421 constructed encoding. However if OpenSSL is compiled with
5422 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
5425 *) Added BIO_vprintf() and BIO_vsnprintf().
5428 *) Added more prefixes to parse for in the the strings written
5429 through a logging bio, to cover all the levels that are available
5430 through syslog. The prefixes are now:
5432 PANIC, EMERG, EMR => LOG_EMERG
5433 ALERT, ALR => LOG_ALERT
5434 CRIT, CRI => LOG_CRIT
5435 ERROR, ERR => LOG_ERR
5436 WARNING, WARN, WAR => LOG_WARNING
5437 NOTICE, NOTE, NOT => LOG_NOTICE
5438 INFO, INF => LOG_INFO
5439 DEBUG, DBG => LOG_DEBUG
5441 and as before, if none of those prefixes are present at the
5442 beginning of the string, LOG_ERR is chosen.
5444 On Win32, the LOG_* levels are mapped according to this:
5446 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
5447 LOG_WARNING => EVENTLOG_WARNING_TYPE
5448 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
5452 *) Made it possible to reconfigure with just the configuration
5453 argument "reconf" or "reconfigure". The command line arguments
5454 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
5455 and are retrieved from there when reconfiguring.
5459 [Assar Westerlund <assar@sics.se>, Richard Levitte]
5461 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
5464 *) The obj_dat.pl script was messing up the sorting of object
5465 names. The reason was that it compared the quoted version
5466 of strings as a result "OCSP" > "OCSP Signing" because
5467 " > SPACE. Changed script to store unquoted versions of
5468 names and add quotes on output. It was also omitting some
5469 names from the lookup table if they were given a default
5470 value (that is if SN is missing it is given the same
5471 value as LN and vice versa), these are now added on the
5472 grounds that if an object has a name we should be able to
5473 look it up. Finally added warning output when duplicate
5474 short or long names are found.
5477 *) Changes needed for Tandem NSK.
5478 [Scott Uroff <scott@xypro.com>]
5480 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
5481 RSA_padding_check_SSLv23(), special padding was never detected
5482 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
5483 version rollback attacks was not effective.
5485 In s23_clnt.c, don't use special rollback-attack detection padding
5486 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
5487 client; similarly, in s23_srvr.c, don't do the rollback check if
5488 SSL 2.0 is the only protocol enabled in the server.
5491 *) Make it possible to get hexdumps of unprintable data with 'openssl
5492 asn1parse'. By implication, the functions ASN1_parse_dump() and
5493 BIO_dump_indent() are added.
5496 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
5497 these print out strings and name structures based on various
5498 flags including RFC2253 support and proper handling of
5499 multibyte characters. Added options to the 'x509' utility
5500 to allow the various flags to be set.
5503 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
5504 Also change the functions X509_cmp_current_time() and
5505 X509_gmtime_adj() work with an ASN1_TIME structure,
5506 this will enable certificates using GeneralizedTime in validity
5507 dates to be checked.
5510 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
5511 negative public key encodings) on by default,
5512 NO_NEG_PUBKEY_BUG can be set to disable it.
5515 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
5516 content octets. An i2c_ASN1_OBJECT is unnecessary because
5517 the encoding can be trivially obtained from the structure.
5520 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
5521 not read locks (CRYPTO_r_[un]lock).
5524 *) A first attempt at creating official support for shared
5525 libraries through configuration. I've kept it so the
5526 default is static libraries only, and the OpenSSL programs
5527 are always statically linked for now, but there are
5528 preparations for dynamic linking in place.
5529 This has been tested on Linux and Tru64.
5532 *) Randomness polling function for Win9x, as described in:
5533 Peter Gutmann, Software Generation of Practically Strong
5537 *) Fix so PRNG is seeded in req if using an already existing
5541 *) New options to smime application. -inform and -outform
5542 allow alternative formats for the S/MIME message including
5543 PEM and DER. The -content option allows the content to be
5544 specified separately. This should allow things like Netscape
5545 form signing output easier to verify.
5548 *) Fix the ASN1 encoding of tags using the 'long form'.
5551 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
5552 STRING types. These convert content octets to and from the
5553 underlying type. The actual tag and length octets are
5554 already assumed to have been read in and checked. These
5555 are needed because all other string types have virtually
5556 identical handling apart from the tag. By having versions
5557 of the ASN1 functions that just operate on content octets
5558 IMPLICIT tagging can be handled properly. It also allows
5559 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
5560 and ASN1_INTEGER are identical apart from the tag.
5563 *) Change the handling of OID objects as follows:
5565 - New object identifiers are inserted in objects.txt, following
5566 the syntax given in objects.README.
5567 - objects.pl is used to process obj_mac.num and create a new
5569 - obj_dat.pl is used to create a new obj_dat.h, using the data in
5572 This is currently kind of a hack, and the perl code in objects.pl
5573 isn't very elegant, but it works as I intended. The simplest way
5574 to check that it worked correctly is to look in obj_dat.h and
5575 check the array nid_objs and make sure the objects haven't moved
5576 around (this is important!). Additions are OK, as well as
5577 consistent name changes.
5580 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
5583 *) Addition of the command line parameter '-rand file' to 'openssl req'.
5584 The given file adds to whatever has already been seeded into the
5585 random pool through the RANDFILE configuration file option or
5586 environment variable, or the default random state file.
5589 *) mkstack.pl now sorts each macro group into lexical order.
5590 Previously the output order depended on the order the files
5591 appeared in the directory, resulting in needless rewriting
5595 *) Patches to make OpenSSL compile under Win32 again. Mostly
5596 work arounds for the VC++ problem that it treats func() as
5597 func(void). Also stripped out the parts of mkdef.pl that
5598 added extra typesafe functions: these no longer exist.
5601 *) Reorganisation of the stack code. The macros are now all
5602 collected in safestack.h . Each macro is defined in terms of
5603 a "stack macro" of the form SKM_<name>(type, a, b). The
5604 DEBUG_SAFESTACK is now handled in terms of function casts,
5605 this has the advantage of retaining type safety without the
5606 use of additional functions. If DEBUG_SAFESTACK is not defined
5607 then the non typesafe macros are used instead. Also modified the
5608 mkstack.pl script to handle the new form. Needs testing to see
5609 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
5610 the default if no major problems. Similar behaviour for ASN1_SET_OF
5611 and PKCS12_STACK_OF.
5614 *) When some versions of IIS use the 'NET' form of private key the
5615 key derivation algorithm is different. Normally MD5(password) is
5616 used as a 128 bit RC4 key. In the modified case
5617 MD5(MD5(password) + "SGCKEYSALT") is used insted. Added some
5618 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
5619 as the old Netscape_RSA functions except they have an additional
5620 'sgckey' parameter which uses the modified algorithm. Also added
5621 an -sgckey command line option to the rsa utility. Thanks to
5622 Adrian Peck <bertie@ncipher.com> for posting details of the modified
5623 algorithm to openssl-dev.
5626 *) The evp_local.h macros were using 'c.##kname' which resulted in
5627 invalid expansion on some systems (SCO 5.0.5 for example).
5628 Corrected to 'c.kname'.
5629 [Phillip Porch <root@theporch.com>]
5631 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
5632 a STACK of email addresses from a certificate or request, these look
5633 in the subject name and the subject alternative name extensions and
5634 omit any duplicate addresses.
5637 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
5638 This makes DSA verification about 2 % faster.
5641 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
5642 (meaning that now 2^5 values will be precomputed, which is only 4 KB
5643 plus overhead for 1024 bit moduli).
5644 This makes exponentiations about 0.5 % faster for 1024 bit
5645 exponents (as measured by "openssl speed rsa2048").
5648 *) Rename memory handling macros to avoid conflicts with other
5650 Malloc => OPENSSL_malloc
5651 Malloc_locked => OPENSSL_malloc_locked
5652 Realloc => OPENSSL_realloc
5653 Free => OPENSSL_free
5656 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
5657 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
5660 *) CygWin32 support.
5661 [John Jarvie <jjarvie@newsguy.com>]
5663 *) The type-safe stack code has been rejigged. It is now only compiled
5664 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
5665 by default all type-specific stack functions are "#define"d back to
5666 standard stack functions. This results in more streamlined output
5667 but retains the type-safety checking possibilities of the original
5671 *) The STACK code has been cleaned up, and certain type declarations
5672 that didn't make a lot of sense have been brought in line. This has
5673 also involved a cleanup of sorts in safestack.h to more correctly
5674 map type-safe stack functions onto their plain stack counterparts.
5675 This work has also resulted in a variety of "const"ifications of
5676 lots of the code, especially "_cmp" operations which should normally
5677 be prototyped with "const" parameters anyway.
5680 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
5681 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
5682 (The PRNG state consists of two parts, the large pool 'state' and 'md',
5683 where all of 'md' is used each time the PRNG is used, but 'state'
5684 is used only indexed by a cyclic counter. As entropy may not be
5685 well distributed from the beginning, 'md' is important as a
5686 chaining variable. However, the output function chains only half
5687 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
5688 all of 'md', and seeding with STATE_SIZE dummy bytes will result
5689 in all of 'state' being rewritten, with the new values depending
5690 on virtually all of 'md'. This overcomes the 80 bit limitation.)
5693 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
5694 the handshake is continued after ssl_verify_cert_chain();
5695 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
5696 can lead to 'unexplainable' connection aborts later.
5697 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
5699 *) Major EVP API cipher revision.
5700 Add hooks for extra EVP features. This allows various cipher
5701 parameters to be set in the EVP interface. Support added for variable
5702 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
5703 setting of RC2 and RC5 parameters.
5705 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
5708 Remove lots of duplicated code from the EVP library. For example *every*
5709 cipher init() function handles the 'iv' in the same way according to the
5710 cipher mode. They also all do nothing if the 'key' parameter is NULL and
5711 for CFB and OFB modes they zero ctx->num.
5713 New functionality allows removal of S/MIME code RC2 hack.
5715 Most of the routines have the same form and so can be declared in terms
5718 By shifting this to the top level EVP_CipherInit() it can be removed from
5719 all individual ciphers. If the cipher wants to handle IVs or keys
5720 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
5723 Change lots of functions like EVP_EncryptUpdate() to now return a
5724 value: although software versions of the algorithms cannot fail
5725 any installed hardware versions can.
5728 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
5729 this option is set, tolerate broken clients that send the negotiated
5730 protocol version number instead of the requested protocol version
5734 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
5735 i.e. non-zero for export ciphersuites, zero otherwise.
5736 Previous versions had this flag inverted, inconsistent with
5737 rsa_tmp_cb (..._TMP_RSA_CB).
5738 [Bodo Moeller; problem reported by Amit Chopra]
5740 *) Add missing DSA library text string. Work around for some IIS
5741 key files with invalid SEQUENCE encoding.
5744 *) Add a document (doc/standards.txt) that list all kinds of standards
5745 and so on that are implemented in OpenSSL.
5748 *) Enhance c_rehash script. Old version would mishandle certificates
5749 with the same subject name hash and wouldn't handle CRLs at all.
5750 Added -fingerprint option to crl utility, to support new c_rehash
5754 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
5757 *) Fix for SSL server purpose checking. Server checking was
5758 rejecting certificates which had extended key usage present
5759 but no ssl client purpose.
5760 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
5762 *) Make PKCS#12 code work with no password. The PKCS#12 spec
5763 is a little unclear about how a blank password is handled.
5764 Since the password in encoded as a BMPString with terminating
5765 double NULL a zero length password would end up as just the
5766 double NULL. However no password at all is different and is
5767 handled differently in the PKCS#12 key generation code. NS
5768 treats a blank password as zero length. MSIE treats it as no
5769 password on export: but it will try both on import. We now do
5770 the same: PKCS12_parse() tries zero length and no password if
5771 the password is set to "" or NULL (NULL is now a valid password:
5772 it wasn't before) as does the pkcs12 application.
5775 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
5776 perror when PEM_read_bio_X509_REQ fails, the error message must
5777 be obtained from the error queue.
5780 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
5781 it in ERR_remove_state if appropriate, and change ERR_get_state
5782 accordingly to avoid race conditions (this is necessary because
5783 thread_hash is no longer constant once set).
5786 *) Bugfix for linux-elf makefile.one.
5789 *) RSA_get_default_method() will now cause a default
5790 RSA_METHOD to be chosen if one doesn't exist already.
5791 Previously this was only set during a call to RSA_new()
5792 or RSA_new_method(NULL) meaning it was possible for
5793 RSA_get_default_method() to return NULL.
5796 *) Added native name translation to the existing DSO code
5797 that will convert (if the flag to do so is set) filenames
5798 that are sufficiently small and have no path information
5799 into a canonical native form. Eg. "blah" converted to
5800 "libblah.so" or "blah.dll" etc.
5803 *) New function ERR_error_string_n(e, buf, len) which is like
5804 ERR_error_string(e, buf), but writes at most 'len' bytes
5805 including the 0 terminator. For ERR_error_string_n, 'buf'
5807 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
5809 *) CONF library reworked to become more general. A new CONF
5810 configuration file reader "class" is implemented as well as a
5811 new functions (NCONF_*, for "New CONF") to handle it. The now
5812 old CONF_* functions are still there, but are reimplemented to
5813 work in terms of the new functions. Also, a set of functions
5814 to handle the internal storage of the configuration data is
5815 provided to make it easier to write new configuration file
5816 reader "classes" (I can definitely see something reading a
5817 configuration file in XML format, for example), called _CONF_*,
5818 or "the configuration storage API"...
5820 The new configuration file reading functions are:
5822 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
5823 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
5825 NCONF_default, NCONF_WIN32
5827 NCONF_dump_fp, NCONF_dump_bio
5829 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
5830 NCONF_new creates a new CONF object. This works in the same way
5831 as other interfaces in OpenSSL, like the BIO interface.
5832 NCONF_dump_* dump the internal storage of the configuration file,
5833 which is useful for debugging. All other functions take the same
5834 arguments as the old CONF_* functions wth the exception of the
5835 first that must be a `CONF *' instead of a `LHASH *'.
5837 To make it easer to use the new classes with the old CONF_* functions,
5838 the function CONF_set_default_method is provided.
5841 *) Add '-tls1' option to 'openssl ciphers', which was already
5842 mentioned in the documentation but had not been implemented.
5843 (This option is not yet really useful because even the additional
5844 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
5847 *) Initial DSO code added into libcrypto for letting OpenSSL (and
5848 OpenSSL-based applications) load shared libraries and bind to
5849 them in a portable way.
5850 [Geoff Thorpe, with contributions from Richard Levitte]
5852 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
5854 *) Make sure _lrotl and _lrotr are only used with MSVC.
5856 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
5857 (the default implementation of RAND_status).
5859 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
5860 to '-clrext' (= clear extensions), as intended and documented.
5861 [Bodo Moeller; inconsistency pointed out by Michael Attili
5864 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
5865 was larger than the MD block size.
5866 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
5868 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
5869 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
5870 using the passed key: if the passed key was a private key the result
5871 of X509_print(), for example, would be to print out all the private key
5875 *) des_quad_cksum() byte order bug fix.
5876 [Ulf Möller, using the problem description in krb4-0.9.7, where
5877 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
5879 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
5881 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
5883 *) For easily testing in shell scripts whether some command
5884 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
5885 returns with exit code 0 iff no command of the given name is available.
5886 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
5887 the output goes to stdout and nothing is printed to stderr.
5888 Additional arguments are always ignored.
5890 Since for each cipher there is a command of the same name,
5891 the 'no-cipher' compilation switches can be tested this way.
5893 ('openssl no-XXX' is not able to detect pseudo-commands such
5894 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
5897 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
5900 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
5901 is set; it will be thrown away anyway because each handshake creates
5903 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
5904 to parameters -- in previous versions (since OpenSSL 0.9.3) the
5905 'default key' from SSL_CTX_set_tmp_dh would always be lost, meanining
5906 you effectivly got SSL_OP_SINGLE_DH_USE when using this macro.
5909 *) New s_client option -ign_eof: EOF at stdin is ignored, and
5910 'Q' and 'R' lose their special meanings (quit/renegotiate).
5911 This is part of what -quiet does; unlike -quiet, -ign_eof
5912 does not suppress any output.
5915 *) Add compatibility options to the purpose and trust code. The
5916 purpose X509_PURPOSE_ANY is "any purpose" which automatically
5917 accepts a certificate or CA, this was the previous behaviour,
5918 with all the associated security issues.
5920 X509_TRUST_COMPAT is the old trust behaviour: only and
5921 automatically trust self signed roots in certificate store. A
5922 new trust setting X509_TRUST_DEFAULT is used to specify that
5923 a purpose has no associated trust setting and it should instead
5924 use the value in the default purpose.
5927 *) Fix the PKCS#8 DSA private key code so it decodes keys again
5928 and fix a memory leak.
5931 *) In util/mkerr.pl (which implements 'make errors'), preserve
5932 reason strings from the previous version of the .c file, as
5933 the default to have only downcase letters (and digits) in
5934 automatically generated reasons codes is not always appropriate.
5937 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
5938 using strerror. Previously, ERR_reason_error_string() returned
5939 library names as reason strings for SYSerr; but SYSerr is a special
5940 case where small numbers are errno values, not library numbers.
5943 *) Add '-dsaparam' option to 'openssl dhparam' application. This
5944 converts DSA parameters into DH parameters. (When creating parameters,
5945 DSA_generate_parameters is used.)
5948 *) Include 'length' (recommended exponent length) in C code generated
5949 by 'openssl dhparam -C'.
5952 *) The second argument to set_label in perlasm was already being used
5953 so couldn't be used as a "file scope" flag. Moved to third argument
5957 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
5958 instead of RAND_bytes for encryption IVs and salts.
5961 *) Include RAND_status() into RAND_METHOD instead of implementing
5962 it only for md_rand.c Otherwise replacing the PRNG by calling
5963 RAND_set_rand_method would be impossible.
5966 *) Don't let DSA_generate_key() enter an infinite loop if the random
5967 number generation fails.
5970 *) New 'rand' application for creating pseudo-random output.
5973 *) Added configuration support for Linux/IA64
5974 [Rolf Haberrecker <rolf@suse.de>]
5976 *) Assembler module support for Mingw32.
5979 *) Shared library support for HPUX (in shlib/).
5980 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
5982 *) Shared library support for Solaris gcc.
5983 [Lutz Behnke <behnke@trustcenter.de>]
5985 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
5987 *) PKCS7_encrypt() was adding text MIME headers twice because they
5988 were added manually and by SMIME_crlf_copy().
5991 *) In bntest.c don't call BN_rand with zero bits argument.
5992 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
5994 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
5995 case was implemented. This caused BN_div_recp() to fail occasionally.
5998 *) Add an optional second argument to the set_label() in the perl
5999 assembly language builder. If this argument exists and is set
6000 to 1 it signals that the assembler should use a symbol whose
6001 scope is the entire file, not just the current function. This
6002 is needed with MASM which uses the format label:: for this scope.
6003 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
6005 *) Change the ASN1 types so they are typedefs by default. Before
6006 almost all types were #define'd to ASN1_STRING which was causing
6007 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
6011 *) Change names of new functions to the new get1/get0 naming
6012 convention: After 'get1', the caller owns a reference count
6013 and has to call ..._free; 'get0' returns a pointer to some
6014 data structure without incrementing reference counters.
6015 (Some of the existing 'get' functions increment a reference
6016 counter, some don't.)
6017 Similarly, 'set1' and 'add1' functions increase reference
6018 counters or duplicate objects.
6021 *) Allow for the possibility of temp RSA key generation failure:
6022 the code used to assume it always worked and crashed on failure.
6025 *) Fix potential buffer overrun problem in BIO_printf().
6026 [Ulf Möller, using public domain code by Patrick Powell; problem
6027 pointed out by David Sacerdote <das33@cornell.edu>]
6029 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
6030 RAND_egd() and RAND_status(). In the command line application,
6031 the EGD socket can be specified like a seed file using RANDFILE
6035 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
6036 Some CAs (e.g. Verisign) distribute certificates in this form.
6039 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
6040 list to exclude them. This means that no special compilation option
6041 is needed to use anonymous DH: it just needs to be included in the
6045 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
6046 EVP_MD_type. The old functionality is available in a new macro called
6047 EVP_MD_md(). Change code that uses it and update docs.
6050 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
6051 where the 'void *' argument is replaced by a function pointer argument.
6052 Previously 'void *' was abused to point to functions, which works on
6053 many platforms, but is not correct. As these functions are usually
6054 called by macros defined in OpenSSL header files, most source code
6055 should work without changes.
6058 *) <openssl/opensslconf.h> (which is created by Configure) now contains
6059 sections with information on -D... compiler switches used for
6060 compiling the library so that applications can see them. To enable
6061 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
6062 must be defined. E.g.,
6063 #define OPENSSL_ALGORITHM_DEFINES
6064 #include <openssl/opensslconf.h>
6065 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
6066 [Richard Levitte, Ulf and Bodo Möller]
6068 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
6072 *) Change the 'other' type in certificate aux info to a STACK_OF
6073 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
6074 the required ASN1 format: arbitrary types determined by an OID.
6077 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
6078 argument to 'req'. This is not because the function is newer or
6079 better than others it just uses the work 'NEW' in the certificate
6080 request header lines. Some software needs this.
6083 *) Reorganise password command line arguments: now passwords can be
6084 obtained from various sources. Delete the PEM_cb function and make
6085 it the default behaviour: i.e. if the callback is NULL and the
6086 usrdata argument is not NULL interpret it as a null terminated pass
6087 phrase. If usrdata and the callback are NULL then the pass phrase
6088 is prompted for as usual.
6091 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
6092 the support is automatically enabled. The resulting binaries will
6093 autodetect the card and use it if present.
6094 [Ben Laurie and Compaq Inc.]
6096 *) Work around for Netscape hang bug. This sends certificate request
6097 and server done in one record. Since this is perfectly legal in the
6098 SSL/TLS protocol it isn't a "bug" option and is on by default. See
6099 the bugs/SSLv3 entry for more info.
6102 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
6105 *) Add -rand argument to smime and pkcs12 applications and read/write
6109 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
6112 *) Add command line password options to the remaining applications.
6115 *) Bug fix for BN_div_recp() for numerators with an even number of
6119 *) More tests in bntest.c, and changed test_bn output.
6122 *) ./config recognizes MacOS X now.
6125 *) Bug fix for BN_div() when the first words of num and divsor are
6126 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
6129 *) Add support for various broken PKCS#8 formats, and command line
6130 options to produce them.
6133 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
6134 get temporary BIGNUMs from a BN_CTX.
6137 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
6141 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
6142 include a #define from the old name to the new. The original intent
6143 was that statically linked binaries could for example just call
6144 SSLeay_add_all_ciphers() to just add ciphers to the table and not
6145 link with digests. This never worked becayse SSLeay_add_all_digests()
6146 and SSLeay_add_all_ciphers() were in the same source file so calling
6147 one would link with the other. They are now in separate source files.
6150 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
6153 *) Use a less unusual form of the Miller-Rabin primality test (it used
6154 a binary algorithm for exponentiation integrated into the Miller-Rabin
6155 loop, our standard modexp algorithms are faster).
6158 *) Support for the EBCDIC character set completed.
6159 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
6161 *) Source code cleanups: use const where appropriate, eliminate casts,
6162 use void * instead of char * in lhash.
6165 *) Bugfix: ssl3_send_server_key_exchange was not restartable
6166 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
6167 this the server could overwrite ephemeral keys that the client
6171 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
6172 using 50 iterations of the Rabin-Miller test.
6174 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
6175 iterations of the Rabin-Miller test as required by the appendix
6176 to FIPS PUB 186[-1]) instead of DSA_is_prime.
6177 As BN_is_prime_fasttest includes trial division, DSA parameter
6178 generation becomes much faster.
6180 This implies a change for the callback functions in DSA_is_prime
6181 and DSA_generate_parameters: The callback function is called once
6182 for each positive witness in the Rabin-Miller test, not just
6183 occasionally in the inner loop; and the parameters to the
6184 callback function now provide an iteration count for the outer
6185 loop rather than for the current invocation of the inner loop.
6186 DSA_generate_parameters additionally can call the callback
6187 function with an 'iteration count' of -1, meaning that a
6188 candidate has passed the trial division test (when q is generated
6189 from an application-provided seed, trial division is skipped).
6192 *) New function BN_is_prime_fasttest that optionally does trial
6193 division before starting the Rabin-Miller test and has
6194 an additional BN_CTX * argument (whereas BN_is_prime always
6195 has to allocate at least one BN_CTX).
6196 'callback(1, -1, cb_arg)' is called when a number has passed the
6197 trial division stage.
6200 *) Fix for bug in CRL encoding. The validity dates weren't being handled
6204 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
6207 *) New function BN_pseudo_rand().
6210 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
6211 bignum version of BN_from_montgomery() with the working code from
6212 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
6216 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
6217 made it impossible to use the same SSL_SESSION data structure in
6218 SSL2 clients in multiple threads.
6221 *) The return value of RAND_load_file() no longer counts bytes obtained
6222 by stat(). RAND_load_file(..., -1) is new and uses the complete file
6223 to seed the PRNG (previously an explicit byte count was required).
6224 [Ulf Möller, Bodo Möller]
6226 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
6227 used (char *) instead of (void *) and had casts all over the place.
6230 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
6233 *) Retain source code compatibility for BN_prime_checks macro:
6234 BN_is_prime(..., BN_prime_checks, ...) now uses
6235 BN_prime_checks_for_size to determine the appropriate number of
6236 Rabin-Miller iterations.
6239 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
6240 DH_CHECK_P_NOT_SAFE_PRIME.
6241 (Check if this is true? OpenPGP calls them "strong".)
6244 *) Merge the functionality of "dh" and "gendh" programs into a new program
6245 "dhparam". The old programs are retained for now but will handle DH keys
6246 (instead of parameters) in future.
6249 *) Make the ciphers, s_server and s_client programs check the return values
6250 when a new cipher list is set.
6253 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
6254 ciphers. Before when the 56bit ciphers were enabled the sorting was
6257 The syntax for the cipher sorting has been extended to support sorting by
6258 cipher-strength (using the strength_bits hard coded in the tables).
6259 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
6261 Fix a bug in the cipher-command parser: when supplying a cipher command
6262 string with an "undefined" symbol (neither command nor alphanumeric
6263 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
6264 an error is flagged.
6266 Due to the strength-sorting extension, the code of the
6267 ssl_create_cipher_list() function was completely rearranged. I hope that
6268 the readability was also increased :-)
6269 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
6271 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
6272 for the first serial number and places 2 in the serial number file. This
6273 avoids problems when the root CA is created with serial number zero and
6274 the first user certificate has the same issuer name and serial number
6278 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
6279 the new code. Add documentation for this stuff.
6282 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
6283 X509_*() to X509at_*() on the grounds that they don't handle X509
6284 structures and behave in an analagous way to the X509v3 functions:
6285 they shouldn't be called directly but wrapper functions should be used
6288 So we also now have some wrapper functions that call the X509at functions
6289 when passed certificate requests. (TO DO: similar things can be done with
6290 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
6291 things. Some of these need some d2i or i2d and print functionality
6292 because they handle more complex structures.)
6295 *) Add missing #ifndefs that caused missing symbols when building libssl
6296 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
6297 NO_RSA in ssl/s2*.c.
6298 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
6300 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
6301 has a return value which indicates the quality of the random data
6302 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
6303 error queue. New function RAND_pseudo_bytes() generates output that is
6304 guaranteed to be unique but not unpredictable. RAND_add is like
6305 RAND_seed, but takes an extra argument for an entropy estimate
6306 (RAND_seed always assumes full entropy).
6309 *) Do more iterations of Rabin-Miller probable prime test (specifically,
6310 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
6311 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
6312 in crypto/bn/bn_prime.c for the complete table). This guarantees a
6313 false-positive rate of at most 2^-80 for random input.
6316 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
6319 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
6320 in the 0.9.5 release), this returns the chain
6321 from an X509_CTX structure with a dup of the stack and all
6322 the X509 reference counts upped: so the stack will exist
6323 after X509_CTX_cleanup() has been called. Modify pkcs12.c
6326 Also make SSL_SESSION_print() print out the verify return
6330 *) Add manpage for the pkcs12 command. Also change the default
6331 behaviour so MAC iteration counts are used unless the new
6332 -nomaciter option is used. This improves file security and
6333 only older versions of MSIE (4.0 for example) need it.
6336 *) Honor the no-xxx Configure options when creating .DEF files.
6339 *) Add PKCS#10 attributes to field table: challengePassword,
6340 unstructuredName and unstructuredAddress. These are taken from
6341 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
6342 international characters are used.
6344 More changes to X509_ATTRIBUTE code: allow the setting of types
6345 based on strings. Remove the 'loc' parameter when adding
6346 attributes because these will be a SET OF encoding which is sorted
6350 *) Initial changes to the 'req' utility to allow request generation
6351 automation. This will allow an application to just generate a template
6352 file containing all the field values and have req construct the
6355 Initial support for X509_ATTRIBUTE handling. Stacks of these are
6356 used all over the place including certificate requests and PKCS#7
6357 structures. They are currently handled manually where necessary with
6358 some primitive wrappers for PKCS#7. The new functions behave in a
6359 manner analogous to the X509 extension functions: they allow
6360 attributes to be looked up by NID and added.
6362 Later something similar to the X509V3 code would be desirable to
6363 automatically handle the encoding, decoding and printing of the
6364 more complex types. The string types like challengePassword can
6365 be handled by the string table functions.
6367 Also modified the multi byte string table handling. Now there is
6368 a 'global mask' which masks out certain types. The table itself
6369 can use the flag STABLE_NO_MASK to ignore the mask setting: this
6370 is useful when for example there is only one permissible type
6371 (as in countryName) and using the mask might result in no valid
6375 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
6376 SSL_get_peer_finished to allow applications to obtain the latest
6377 Finished messages sent to the peer or expected from the peer,
6378 respectively. (SSL_get_peer_finished is usually the Finished message
6379 actually received from the peer, otherwise the protocol will be aborted.)
6381 As the Finished message are message digests of the complete handshake
6382 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
6383 be used for external authentication procedures when the authentication
6384 provided by SSL/TLS is not desired or is not enough.
6387 *) Enhanced support for Alpha Linux is added. Now ./config checks if
6388 the host supports BWX extension and if Compaq C is present on the
6389 $PATH. Just exploiting of the BWX extension results in 20-30%
6390 performance kick for some algorithms, e.g. DES and RC4 to mention
6391 a couple. Compaq C in turn generates ~20% faster code for MD5 and
6395 *) Add support for MS "fast SGC". This is arguably a violation of the
6396 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
6397 weak crypto and after checking the certificate is SGC a second one
6398 with strong crypto. MS SGC stops the first handshake after receiving
6399 the server certificate message and sends a second client hello. Since
6400 a server will typically do all the time consuming operations before
6401 expecting any further messages from the client (server key exchange
6402 is the most expensive) there is little difference between the two.
6404 To get OpenSSL to support MS SGC we have to permit a second client
6405 hello message after we have sent server done. In addition we have to
6406 reset the MAC if we do get this second client hello.
6409 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
6410 if a DER encoded private key is RSA or DSA traditional format. Changed
6411 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
6412 format DER encoded private key. Newer code should use PKCS#8 format which
6413 has the key type encoded in the ASN1 structure. Added DER private key
6414 support to pkcs8 application.
6417 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
6418 ciphersuites has been selected (as required by the SSL 3/TLS 1
6419 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
6420 is set, we interpret this as a request to violate the specification
6421 (the worst that can happen is a handshake failure, and 'correct'
6422 behaviour would result in a handshake failure anyway).
6425 *) In SSL_CTX_add_session, take into account that there might be multiple
6426 SSL_SESSION structures with the same session ID (e.g. when two threads
6427 concurrently obtain them from an external cache).
6428 The internal cache can handle only one SSL_SESSION with a given ID,
6429 so if there's a conflict, we now throw out the old one to achieve
6433 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
6434 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
6435 some routines that use cipher OIDs: some ciphers do not have OIDs
6436 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
6440 *) Simplify the trust setting structure and code. Now we just have
6441 two sequences of OIDs for trusted and rejected settings. These will
6442 typically have values the same as the extended key usage extension
6443 and any application specific purposes.
6445 The trust checking code now has a default behaviour: it will just
6446 check for an object with the same NID as the passed id. Functions can
6447 be provided to override either the default behaviour or the behaviour
6448 for a given id. SSL client, server and email already have functions
6449 in place for compatibility: they check the NID and also return "trusted"
6450 if the certificate is self signed.
6453 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
6454 traditional format into an EVP_PKEY structure.
6457 *) Add a password callback function PEM_cb() which either prompts for
6458 a password if usr_data is NULL or otherwise assumes it is a null
6459 terminated password. Allow passwords to be passed on command line
6460 environment or config files in a few more utilities.
6463 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
6464 keys. Add some short names for PKCS#8 PBE algorithms and allow them
6465 to be specified on the command line for the pkcs8 and pkcs12 utilities.
6466 Update documentation.
6469 *) Support for ASN1 "NULL" type. This could be handled before by using
6470 ASN1_TYPE but there wasn't any function that would try to read a NULL
6471 and produce an error if it couldn't. For compatibility we also have
6472 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
6473 don't allocate anything because they don't need to.
6476 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
6478 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
6480 *) Rebuild of the memory allocation routines used by OpenSSL code and
6481 possibly others as well. The purpose is to make an interface that
6482 provide hooks so anyone can build a separate set of allocation and
6483 deallocation routines to be used by OpenSSL, for example memory
6484 pool implementations, or something else, which was previously hard
6485 since Malloc(), Realloc() and Free() were defined as macros having
6486 the values malloc, realloc and free, respectively (except for Win32
6487 compilations). The same is provided for memory debugging code.
6488 OpenSSL already comes with functionality to find memory leaks, but
6489 this gives people a chance to debug other memory problems.
6491 With these changes, a new set of functions and macros have appeared:
6493 CRYPTO_set_mem_debug_functions() [F]
6494 CRYPTO_get_mem_debug_functions() [F]
6495 CRYPTO_dbg_set_options() [F]
6496 CRYPTO_dbg_get_options() [F]
6497 CRYPTO_malloc_debug_init() [M]
6499 The memory debug functions are NULL by default, unless the library
6500 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
6501 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
6502 gives the standard debugging functions that come with OpenSSL) or
6503 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
6504 provided by the library user) must be used. When the standard
6505 debugging functions are used, CRYPTO_dbg_set_options can be used to
6506 request additional information:
6507 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
6508 the CRYPTO_MDEBUG_xxx macro when compiling the library.
6510 Also, things like CRYPTO_set_mem_functions will always give the
6511 expected result (the new set of functions is used for allocation
6512 and deallocation) at all times, regardless of platform and compiler
6515 To finish it up, some functions that were never use in any other
6516 way than through macros have a new API and new semantic:
6519 CRYPTO_dbg_realloc()
6522 All macros of value have retained their old syntax.
6523 [Richard Levitte and Bodo Moeller]
6525 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
6526 ordering of SMIMECapabilities wasn't in "strength order" and there
6527 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
6531 *) Some ASN1 types with illegal zero length encoding (INTEGER,
6532 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
6533 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
6535 *) Merge in my S/MIME library for OpenSSL. This provides a simple
6536 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
6537 functionality to handle multipart/signed properly) and a utility
6538 called 'smime' to call all this stuff. This is based on code I
6539 originally wrote for Celo who have kindly allowed it to be
6540 included in OpenSSL.
6543 *) Add variants des_set_key_checked and des_set_key_unchecked of
6544 des_set_key (aka des_key_sched). Global variable des_check_key
6545 decides which of these is called by des_set_key; this way
6546 des_check_key behaves as it always did, but applications and
6547 the library itself, which was buggy for des_check_key == 1,
6548 have a cleaner way to pick the version they need.
6551 *) New function PKCS12_newpass() which changes the password of a
6555 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
6556 dynamic mix. In both cases the ids can be used as an index into the
6557 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
6558 functions so they accept a list of the field values and the
6559 application doesn't need to directly manipulate the X509_TRUST
6563 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
6567 *) Modify the way the V3 extension code looks up extensions. This now
6568 works in a similar way to the object code: we have some "standard"
6569 extensions in a static table which is searched with OBJ_bsearch()
6570 and the application can add dynamic ones if needed. The file
6571 crypto/x509v3/ext_dat.h now has the info: this file needs to be
6572 updated whenever a new extension is added to the core code and kept
6573 in ext_nid order. There is a simple program 'tabtest.c' which checks
6574 this. New extensions are not added too often so this file can readily
6575 be maintained manually.
6577 There are two big advantages in doing things this way. The extensions
6578 can be looked up immediately and no longer need to be "added" using
6579 X509V3_add_standard_extensions(): this function now does nothing.
6580 [Side note: I get *lots* of email saying the extension code doesn't
6581 work because people forget to call this function]
6582 Also no dynamic allocation is done unless new extensions are added:
6583 so if we don't add custom extensions there is no need to call
6584 X509V3_EXT_cleanup().
6587 *) Modify enc utility's salting as follows: make salting the default. Add a
6588 magic header, so unsalted files fail gracefully instead of just decrypting
6589 to garbage. This is because not salting is a big security hole, so people
6590 should be discouraged from doing it.
6593 *) Fixes and enhancements to the 'x509' utility. It allowed a message
6594 digest to be passed on the command line but it only used this
6595 parameter when signing a certificate. Modified so all relevant
6596 operations are affected by the digest parameter including the
6597 -fingerprint and -x509toreq options. Also -x509toreq choked if a
6598 DSA key was used because it didn't fix the digest.
6601 *) Initial certificate chain verify code. Currently tests the untrusted
6602 certificates for consistency with the verify purpose (which is set
6603 when the X509_STORE_CTX structure is set up) and checks the pathlength.
6605 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
6606 this is because it will reject chains with invalid extensions whereas
6607 every previous version of OpenSSL and SSLeay made no checks at all.
6609 Trust code: checks the root CA for the relevant trust settings. Trust
6610 settings have an initial value consistent with the verify purpose: e.g.
6611 if the verify purpose is for SSL client use it expects the CA to be
6612 trusted for SSL client use. However the default value can be changed to
6613 permit custom trust settings: one example of this would be to only trust
6614 certificates from a specific "secure" set of CAs.
6616 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
6617 which should be used for version portability: especially since the
6618 verify structure is likely to change more often now.
6620 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
6621 to set them. If not set then assume SSL clients will verify SSL servers
6624 Two new options to the verify program: -untrusted allows a set of
6625 untrusted certificates to be passed in and -purpose which sets the
6626 intended purpose of the certificate. If a purpose is set then the
6627 new chain verify code is used to check extension consistency.
6630 *) Support for the authority information access extension.
6633 *) Modify RSA and DSA PEM read routines to transparently handle
6634 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
6635 public keys in a format compatible with certificate
6636 SubjectPublicKeyInfo structures. Unfortunately there were already
6637 functions called *_PublicKey_* which used various odd formats so
6638 these are retained for compatibility: however the DSA variants were
6639 never in a public release so they have been deleted. Changed dsa/rsa
6640 utilities to handle the new format: note no releases ever handled public
6641 keys so we should be OK.
6643 The primary motivation for this change is to avoid the same fiasco
6644 that dogs private keys: there are several incompatible private key
6645 formats some of which are standard and some OpenSSL specific and
6646 require various evil hacks to allow partial transparent handling and
6647 even then it doesn't work with DER formats. Given the option anything
6648 other than PKCS#8 should be dumped: but the other formats have to
6649 stay in the name of compatibility.
6651 With public keys and the benefit of hindsight one standard format
6652 is used which works with EVP_PKEY, RSA or DSA structures: though
6653 it clearly returns an error if you try to read the wrong kind of key.
6655 Added a -pubkey option to the 'x509' utility to output the public key.
6656 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
6657 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
6658 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
6659 that do the same as the EVP_PKEY_assign_*() except they up the
6660 reference count of the added key (they don't "swallow" the
6664 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
6665 CRLs would fail if the file contained no certificates or no CRLs:
6666 added a new function to read in both types and return the number
6667 read: this means that if none are read it will be an error. The
6668 DER versions of the certificate and CRL reader would always fail
6669 because it isn't possible to mix certificates and CRLs in DER format
6670 without choking one or the other routine. Changed this to just read
6671 a certificate: this is the best we can do. Also modified the code
6672 in apps/verify.c to take notice of return codes: it was previously
6673 attempting to read in certificates from NULL pointers and ignoring
6674 any errors: this is one reason why the cert and CRL reader seemed
6675 to work. It doesn't check return codes from the default certificate
6676 routines: these may well fail if the certificates aren't installed.
6679 *) Code to support otherName option in GeneralName.
6682 *) First update to verify code. Change the verify utility
6683 so it warns if it is passed a self signed certificate:
6684 for consistency with the normal behaviour. X509_verify
6685 has been modified to it will now verify a self signed
6686 certificate if *exactly* the same certificate appears
6687 in the store: it was previously impossible to trust a
6688 single self signed certificate. This means that:
6689 openssl verify ss.pem
6690 now gives a warning about a self signed certificate but
6691 openssl verify -CAfile ss.pem ss.pem
6695 *) For servers, store verify_result in SSL_SESSION data structure
6696 (and add it to external session representation).
6697 This is needed when client certificate verifications fails,
6698 but an application-provided verification callback (set by
6699 SSL_CTX_set_cert_verify_callback) allows accepting the session
6700 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
6701 but returns 1): When the session is reused, we have to set
6702 ssl->verify_result to the appropriate error code to avoid
6704 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
6706 *) Fix a bug in the new PKCS#7 code: it didn't consider the
6707 case in PKCS7_dataInit() where the signed PKCS7 structure
6708 didn't contain any existing data because it was being created.
6709 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
6711 *) Add a salt to the key derivation routines in enc.c. This
6712 forms the first 8 bytes of the encrypted file. Also add a
6713 -S option to allow a salt to be input on the command line.
6716 *) New function X509_cmp(). Oddly enough there wasn't a function
6717 to compare two certificates. We do this by working out the SHA1
6718 hash and comparing that. X509_cmp() will be needed by the trust
6722 *) SSL_get1_session() is like SSL_get_session(), but increments
6723 the reference count in the SSL_SESSION returned.
6724 [Geoff Thorpe <geoff@eu.c2.net>]
6726 *) Fix for 'req': it was adding a null to request attributes.
6727 Also change the X509_LOOKUP and X509_INFO code to handle
6728 certificate auxiliary information.
6731 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
6735 *) Add the possibility to add extra information to the memory leak
6736 detecting output, to form tracebacks, showing from where each
6737 allocation was originated: CRYPTO_push_info("constant string") adds
6738 the string plus current file name and line number to a per-thread
6739 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
6740 is like calling CYRPTO_pop_info() until the stack is empty.
6741 Also updated memory leak detection code to be multi-thread-safe.
6744 *) Add options -text and -noout to pkcs7 utility and delete the
6745 encryption options which never did anything. Update docs.
6748 *) Add options to some of the utilities to allow the pass phrase
6749 to be included on either the command line (not recommended on
6750 OSes like Unix) or read from the environment. Update the
6751 manpages and fix a few bugs.
6754 *) Add a few manpages for some of the openssl commands.
6757 *) Fix the -revoke option in ca. It was freeing up memory twice,
6758 leaking and not finding already revoked certificates.
6761 *) Extensive changes to support certificate auxiliary information.
6762 This involves the use of X509_CERT_AUX structure and X509_AUX
6763 functions. An X509_AUX function such as PEM_read_X509_AUX()
6764 can still read in a certificate file in the usual way but it
6765 will also read in any additional "auxiliary information". By
6766 doing things this way a fair degree of compatibility can be
6767 retained: existing certificates can have this information added
6768 using the new 'x509' options.
6770 Current auxiliary information includes an "alias" and some trust
6771 settings. The trust settings will ultimately be used in enhanced
6772 certificate chain verification routines: currently a certificate
6773 can only be trusted if it is self signed and then it is trusted
6777 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
6778 The problem was that one of the replacement routines had not been working
6779 since SSLeay releases. For now the offending routine has been replaced
6780 with non-optimised assembler. Even so, this now gives around 95%
6781 performance improvement for 1024 bit RSA signs.
6784 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
6785 handling. Most clients have the effective key size in bits equal to
6786 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
6787 A few however don't do this and instead use the size of the decrypted key
6788 to determine the RC2 key length and the AlgorithmIdentifier to determine
6789 the effective key length. In this case the effective key length can still
6790 be 40 bits but the key length can be 168 bits for example. This is fixed
6791 by manually forcing an RC2 key into the EVP_PKEY structure because the
6792 EVP code can't currently handle unusual RC2 key sizes: it always assumes
6793 the key length and effective key length are equal.
6796 *) Add a bunch of functions that should simplify the creation of
6797 X509_NAME structures. Now you should be able to do:
6798 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
6799 and have it automatically work out the correct field type and fill in
6800 the structures. The more adventurous can try:
6801 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
6802 and it will (hopefully) work out the correct multibyte encoding.
6805 *) Change the 'req' utility to use the new field handling and multibyte
6806 copy routines. Before the DN field creation was handled in an ad hoc
6807 way in req, ca, and x509 which was rather broken and didn't support
6808 BMPStrings or UTF8Strings. Since some software doesn't implement
6809 BMPStrings or UTF8Strings yet, they can be enabled using the config file
6810 using the dirstring_type option. See the new comment in the default
6811 openssl.cnf for more info.
6814 *) Make crypto/rand/md_rand.c more robust:
6815 - Assure unique random numbers after fork().
6816 - Make sure that concurrent threads access the global counter and
6817 md serializably so that we never lose entropy in them
6818 or use exactly the same state in multiple threads.
6819 Access to the large state is not always serializable because
6820 the additional locking could be a performance killer, and
6821 md should be large enough anyway.
6824 *) New file apps/app_rand.c with commonly needed functionality
6825 for handling the random seed file.
6827 Use the random seed file in some applications that previously did not:
6829 dsaparam -genkey (which also ignored its '-rand' option),
6832 x509 (when signing).
6833 Except on systems with /dev/urandom, it is crucial to have a random
6834 seed file at least for key creation, DSA signing, and for DH exchanges;
6835 for RSA signatures we could do without one.
6837 gendh and gendsa (unlike genrsa) used to read only the first byte
6838 of each file listed in the '-rand' option. The function as previously
6839 found in genrsa is now in app_rand.c and is used by all programs
6840 that support '-rand'.
6843 *) In RAND_write_file, use mode 0600 for creating files;
6844 don't just chmod when it may be too late.
6847 *) Report an error from X509_STORE_load_locations
6848 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
6851 *) New function ASN1_mbstring_copy() this copies a string in either
6852 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
6853 into an ASN1_STRING type. A mask of permissible types is passed
6854 and it chooses the "minimal" type to use or an error if not type
6858 *) Add function equivalents to the various macros in asn1.h. The old
6859 macros are retained with an M_ prefix. Code inside the library can
6860 use the M_ macros. External code (including the openssl utility)
6861 should *NOT* in order to be "shared library friendly".
6864 *) Add various functions that can check a certificate's extensions
6865 to see if it usable for various purposes such as SSL client,
6866 server or S/MIME and CAs of these types. This is currently
6867 VERY EXPERIMENTAL but will ultimately be used for certificate chain
6868 verification. Also added a -purpose flag to x509 utility to
6869 print out all the purposes.
6872 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
6876 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
6877 for, obtain and decode and extension and obtain its critical flag.
6878 This allows all the necessary extension code to be handled in a
6879 single function call.
6882 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
6883 platforms. See crypto/rc4/rc4_enc.c for further details.
6886 *) New -noout option to asn1parse. This causes no output to be produced
6887 its main use is when combined with -strparse and -out to extract data
6888 from a file (which may not be in ASN.1 format).
6891 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
6892 when producing the local key id.
6893 [Richard Levitte <levitte@stacken.kth.se>]
6895 *) New option -dhparam in s_server. This allows a DH parameter file to be
6896 stated explicitly. If it is not stated then it tries the first server
6897 certificate file. The previous behaviour hard coded the filename
6901 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
6902 a public key to be input or output. For example:
6903 openssl rsa -in key.pem -pubout -out pubkey.pem
6904 Also added necessary DSA public key functions to handle this.
6907 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
6908 in the message. This was handled by allowing
6909 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
6910 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
6912 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
6913 to the end of the strings whereas this didn't. This would cause problems
6914 if strings read with d2i_ASN1_bytes() were later modified.
6915 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
6917 *) Fix for base64 decode bug. When a base64 bio reads only one line of
6918 data and it contains EOF it will end up returning an error. This is
6919 caused by input 46 bytes long. The cause is due to the way base64
6920 BIOs find the start of base64 encoded data. They do this by trying a
6921 trial decode on each line until they find one that works. When they
6922 do a flag is set and it starts again knowing it can pass all the
6923 data directly through the decoder. Unfortunately it doesn't reset
6924 the context it uses. This means that if EOF is reached an attempt
6925 is made to pass two EOFs through the context and this causes the
6926 resulting error. This can also cause other problems as well. As is
6927 usual with these problems it takes *ages* to find and the fix is
6928 trivial: move one line.
6929 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
6931 *) Ugly workaround to get s_client and s_server working under Windows. The
6932 old code wouldn't work because it needed to select() on sockets and the
6933 tty (for keypresses and to see if data could be written). Win32 only
6934 supports select() on sockets so we select() with a 1s timeout on the
6935 sockets and then see if any characters are waiting to be read, if none
6936 are present then we retry, we also assume we can always write data to
6937 the tty. This isn't nice because the code then blocks until we've
6938 received a complete line of data and it is effectively polling the
6939 keyboard at 1s intervals: however it's quite a bit better than not
6940 working at all :-) A dedicated Windows application might handle this
6941 with an event loop for example.
6944 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
6945 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
6946 will be called when RSA_sign() and RSA_verify() are used. This is useful
6947 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
6948 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
6949 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
6950 This necessitated the support of an extra signature type NID_md5_sha1
6951 for SSL signatures and modifications to the SSL library to use it instead
6952 of calling RSA_public_decrypt() and RSA_private_encrypt().
6955 *) Add new -verify -CAfile and -CApath options to the crl program, these
6956 will lookup a CRL issuers certificate and verify the signature in a
6957 similar way to the verify program. Tidy up the crl program so it
6958 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
6959 less strict. It will now permit CRL extensions even if it is not
6960 a V2 CRL: this will allow it to tolerate some broken CRLs.
6963 *) Initialize all non-automatic variables each time one of the openssl
6964 sub-programs is started (this is necessary as they may be started
6965 multiple times from the "OpenSSL>" prompt).
6966 [Lennart Bang, Bodo Moeller]
6968 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
6969 removing all other RSA functionality (this is what NO_RSA does). This
6970 is so (for example) those in the US can disable those operations covered
6971 by the RSA patent while allowing storage and parsing of RSA keys and RSA
6975 *) Non-copying interface to BIO pairs.
6976 (still largely untested)
6979 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
6980 ASCII string. This was handled independently in various places before.
6983 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
6984 UTF8 strings a character at a time.
6987 *) Use client_version from client hello to select the protocol
6988 (s23_srvr.c) and for RSA client key exchange verification
6989 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
6992 *) Add various utility functions to handle SPKACs, these were previously
6993 handled by poking round in the structure internals. Added new function
6994 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
6995 print, verify and generate SPKACs. Based on an original idea from
6996 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
6999 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
7002 *) Allow the config file extension section to be overwritten on the
7003 command line. Based on an original idea from Massimiliano Pala
7004 <madwolf@comune.modena.it>. The new option is called -extensions
7005 and can be applied to ca, req and x509. Also -reqexts to override
7006 the request extensions in req and -crlexts to override the crl extensions
7010 *) Add new feature to the SPKAC handling in ca. Now you can include
7011 the same field multiple times by preceding it by "XXXX." for example:
7014 this is the same syntax as used in the req config file.
7017 *) Allow certificate extensions to be added to certificate requests. These
7018 are specified in a 'req_extensions' option of the req section of the
7019 config file. They can be printed out with the -text option to req but
7020 are otherwise ignored at present.
7023 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
7024 data read consists of only the final block it would not decrypted because
7025 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
7026 A misplaced 'break' also meant the decrypted final block might not be
7027 copied until the next read.
7030 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
7031 a few extra parameters to the DH structure: these will be useful if
7032 for example we want the value of 'q' or implement X9.42 DH.
7035 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
7036 provides hooks that allow the default DSA functions or functions on a
7037 "per key" basis to be replaced. This allows hardware acceleration and
7038 hardware key storage to be handled without major modification to the
7039 library. Also added low level modexp hooks and CRYPTO_EX structure and
7040 associated functions.
7043 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
7044 as "read only": it can't be written to and the buffer it points to will
7045 not be freed. Reading from a read only BIO is much more efficient than
7046 a normal memory BIO. This was added because there are several times when
7047 an area of memory needs to be read from a BIO. The previous method was
7048 to create a memory BIO and write the data to it, this results in two
7049 copies of the data and an O(n^2) reading algorithm. There is a new
7050 function BIO_new_mem_buf() which creates a read only memory BIO from
7051 an area of memory. Also modified the PKCS#7 routines to use read only
7055 *) Bugfix: ssl23_get_client_hello did not work properly when called in
7056 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
7057 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
7058 but a retry condition occured while trying to read the rest.
7061 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
7062 NID_pkcs7_encrypted by default: this was wrong since this should almost
7063 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
7064 the encrypted data type: this is a more sensible place to put it and it
7065 allows the PKCS#12 code to be tidied up that duplicated this
7069 *) Changed obj_dat.pl script so it takes its input and output files on
7070 the command line. This should avoid shell escape redirection problems
7074 *) Initial support for certificate extension requests, these are included
7075 in things like Xenroll certificate requests. Included functions to allow
7076 extensions to be obtained and added.
7079 *) -crlf option to s_client and s_server for sending newlines as
7080 CRLF (as required by many protocols).
7083 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
7085 *) Install libRSAglue.a when OpenSSL is built with RSAref.
7086 [Ralf S. Engelschall]
7088 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
7089 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
7091 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
7095 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
7096 DH parameters/keys (q is lost during that conversion, but the resulting
7097 DH parameters contain its length).
7099 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
7100 much faster than DH_generate_parameters (which creates parameters
7101 where p = 2*q + 1), and also the smaller q makes DH computations
7102 much more efficient (160-bit exponentiation instead of 1024-bit
7103 exponentiation); so this provides a convenient way to support DHE
7104 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
7105 utter importance to use
7106 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7108 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
7109 when such DH parameters are used, because otherwise small subgroup
7110 attacks may become possible!
7113 *) Avoid memory leak in i2d_DHparams.
7116 *) Allow the -k option to be used more than once in the enc program:
7117 this allows the same encrypted message to be read by multiple recipients.
7120 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
7121 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
7122 it will always use the numerical form of the OID, even if it has a short
7126 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
7127 method only got called if p,q,dmp1,dmq1,iqmp components were present,
7128 otherwise bn_mod_exp was called. In the case of hardware keys for example
7129 no private key components need be present and it might store extra data
7130 in the RSA structure, which cannot be accessed from bn_mod_exp.
7131 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
7132 private key operations.
7135 *) Added support for SPARC Linux.
7138 *) pem_password_cb function type incompatibly changed from
7139 typedef int pem_password_cb(char *buf, int size, int rwflag);
7141 ....(char *buf, int size, int rwflag, void *userdata);
7142 so that applications can pass data to their callbacks:
7143 The PEM[_ASN1]_{read,write}... functions and macros now take an
7144 additional void * argument, which is just handed through whenever
7145 the password callback is called.
7146 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
7148 New function SSL_CTX_set_default_passwd_cb_userdata.
7150 Compatibility note: As many C implementations push function arguments
7151 onto the stack in reverse order, the new library version is likely to
7152 interoperate with programs that have been compiled with the old
7153 pem_password_cb definition (PEM_whatever takes some data that
7154 happens to be on the stack as its last argument, and the callback
7155 just ignores this garbage); but there is no guarantee whatsoever that
7158 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
7159 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
7160 problems not only on Windows, but also on some Unix platforms.
7161 To avoid problematic command lines, these definitions are now in an
7162 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
7163 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
7166 *) MIPS III/IV assembler module is reimplemented.
7169 *) More DES library cleanups: remove references to srand/rand and
7170 delete an unused file.
7173 *) Add support for the the free Netwide assembler (NASM) under Win32,
7174 since not many people have MASM (ml) and it can be hard to obtain.
7175 This is currently experimental but it seems to work OK and pass all
7176 the tests. Check out INSTALL.W32 for info.
7179 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
7180 without temporary keys kept an extra copy of the server key,
7181 and connections with temporary keys did not free everything in case
7185 *) New function RSA_check_key and new openssl rsa option -check
7186 for verifying the consistency of RSA keys.
7187 [Ulf Moeller, Bodo Moeller]
7189 *) Various changes to make Win32 compile work:
7190 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
7191 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
7192 comparison" warnings.
7193 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
7196 *) Add a debugging option to PKCS#5 v2 key generation function: when
7197 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
7198 derived keys are printed to stderr.
7201 *) Copy the flags in ASN1_STRING_dup().
7202 [Roman E. Pavlov <pre@mo.msk.ru>]
7204 *) The x509 application mishandled signing requests containing DSA
7205 keys when the signing key was also DSA and the parameters didn't match.
7207 It was supposed to omit the parameters when they matched the signing key:
7208 the verifying software was then supposed to automatically use the CA's
7209 parameters if they were absent from the end user certificate.
7211 Omitting parameters is no longer recommended. The test was also
7212 the wrong way round! This was probably due to unusual behaviour in
7213 EVP_cmp_parameters() which returns 1 if the parameters match.
7214 This meant that parameters were omitted when they *didn't* match and
7215 the certificate was useless. Certificates signed with 'ca' didn't have
7217 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
7219 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
7220 The interface is as follows:
7221 Applications can use
7222 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
7223 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
7224 "off" is now the default.
7225 The library internally uses
7226 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
7227 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
7228 to disable memory-checking temporarily.
7230 Some inconsistent states that previously were possible (and were
7231 even the default) are now avoided.
7233 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
7234 with each memory chunk allocated; this is occasionally more helpful
7235 than just having a counter.
7237 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
7239 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
7243 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
7244 which largely parallels "options", but is for changing API behaviour,
7245 whereas "options" are about protocol behaviour.
7246 Initial "mode" flags are:
7248 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
7249 a single record has been written.
7250 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
7251 retries use the same buffer location.
7252 (But all of the contents must be
7256 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
7259 *) Fix problems with no-hmac etc.
7260 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
7262 *) New functions RSA_get_default_method(), RSA_set_method() and
7263 RSA_get_method(). These allows replacement of RSA_METHODs without having
7264 to mess around with the internals of an RSA structure.
7267 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
7268 Also really enable memory leak checks in openssl.c and in some
7270 [Chad C. Mulligan, Bodo Moeller]
7272 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
7273 up the length of negative integers. This has now been simplified to just
7274 store the length when it is first determined and use it later, rather
7275 than trying to keep track of where data is copied and updating it to
7277 [Steve Henson, reported by Brien Wheeler
7278 <bwheeler@authentica-security.com>]
7280 *) Add a new function PKCS7_signatureVerify. This allows the verification
7281 of a PKCS#7 signature but with the signing certificate passed to the
7282 function itself. This contrasts with PKCS7_dataVerify which assumes the
7283 certificate is present in the PKCS#7 structure. This isn't always the
7284 case: certificates can be omitted from a PKCS#7 structure and be
7285 distributed by "out of band" means (such as a certificate database).
7288 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
7289 function prototypes in pem.h, also change util/mkdef.pl to add the
7290 necessary function names.
7293 *) mk1mf.pl (used by Windows builds) did not properly read the
7294 options set by Configure in the top level Makefile, and Configure
7295 was not even able to write more than one option correctly.
7296 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
7299 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
7300 file to be loaded from a BIO or FILE pointer. The BIO version will
7301 for example allow memory BIOs to contain config info.
7304 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
7305 Whoever hopes to achieve shared-library compatibility across versions
7306 must use this, not the compile-time macro.
7307 (Exercise 0.9.4: Which is the minimum library version required by
7309 Note: All this applies only to multi-threaded programs, others don't
7313 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
7314 through a BIO pair triggered the default case, i.e.
7315 SSLerr(...,SSL_R_UNKNOWN_STATE).
7318 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
7319 can use the SSL library even if none of the specific BIOs is
7323 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
7324 for the encoded length.
7325 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
7327 *) Add initial documentation of the X509V3 functions.
7330 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
7331 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
7332 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
7333 secure PKCS#8 private key format with a high iteration count.
7336 *) Fix determination of Perl interpreter: A perl or perl5
7337 _directory_ in $PATH was also accepted as the interpreter.
7338 [Ralf S. Engelschall]
7340 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
7341 wrong with it but it was very old and did things like calling
7342 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
7346 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
7347 to use the new extension code.
7350 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
7351 with macros. This should make it easier to change their form, add extra
7352 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
7356 *) Add to configuration table a new entry that can specify an alternative
7357 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
7358 according to Mark Crispin <MRC@Panda.COM>.
7362 *) DES CBC did not update the IV. Weird.
7365 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
7366 Changing the behaviour of the former might break existing programs --
7367 where IV updating is needed, des_ncbc_encrypt can be used.
7370 *) When bntest is run from "make test" it drives bc to check its
7371 calculations, as well as internally checking them. If an internal check
7372 fails, it needs to cause bc to give a non-zero result or make test carries
7373 on without noticing the failure. Fixed.
7376 *) DES library cleanups.
7379 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
7380 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
7381 ciphers. NOTE: although the key derivation function has been verified
7382 against some published test vectors it has not been extensively tested
7383 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
7387 *) Instead of "mkdir -p", which is not fully portable, use new
7388 Perl script "util/mkdir-p.pl".
7391 *) Rewrite the way password based encryption (PBE) is handled. It used to
7392 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
7393 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
7394 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
7395 the 'parameter' field of the AlgorithmIdentifier is passed to the
7396 underlying key generation function so it must do its own ASN1 parsing.
7397 This has also changed the EVP_PBE_CipherInit() function which now has a
7398 'parameter' argument instead of literal salt and iteration count values
7399 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
7402 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
7403 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
7404 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
7405 KEY" because this clashed with PKCS#8 unencrypted string. Since this
7406 value was just used as a "magic string" and not used directly its
7407 value doesn't matter.
7410 *) Introduce some semblance of const correctness to BN. Shame C doesn't
7414 *) "linux-sparc64" configuration (ultrapenguin).
7415 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
7416 "linux-sparc" configuration.
7417 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
7419 *) config now generates no-xxx options for missing ciphers.
7422 *) Support the EBCDIC character set (work in progress).
7423 File ebcdic.c not yet included because it has a different license.
7424 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7426 *) Support BS2000/OSD-POSIX.
7427 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
7429 *) Make callbacks for key generation use void * instead of char *.
7432 *) Make S/MIME samples compile (not yet tested).
7435 *) Additional typesafe stacks.
7438 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
7442 Changes between 0.9.3 and 0.9.3a [29 May 1999]
7444 *) New configuration variant "sco5-gcc".
7446 *) Updated some demos.
7447 [Sean O Riordain, Wade Scholine]
7449 *) Add missing BIO_free at exit of pkcs12 application.
7452 *) Fix memory leak in conf.c.
7455 *) Updates for Win32 to assembler version of MD5.
7458 *) Set #! path to perl in apps/der_chop to where we found it
7459 instead of using a fixed path.
7462 *) SHA library changes for irix64-mips4-cc.
7465 *) Improvements for VMS support.
7469 Changes between 0.9.2b and 0.9.3 [24 May 1999]
7471 *) Bignum library bug fix. IRIX 6 passes "make test" now!
7472 This also avoids the problems with SC4.2 and unpatched SC5.
7473 [Andy Polyakov <appro@fy.chalmers.se>]
7475 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
7476 These are required because of the typesafe stack would otherwise break
7477 existing code. If old code used a structure member which used to be STACK
7478 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
7479 sk_num or sk_value it would produce an error because the num, data members
7480 are not present in STACK_OF. Now it just produces a warning. sk_set
7481 replaces the old method of assigning a value to sk_value
7482 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
7483 that does this will no longer work (and should use sk_set instead) but
7484 this could be regarded as a "questionable" behaviour anyway.
7487 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
7488 correctly handle encrypted S/MIME data.
7491 *) Change type of various DES function arguments from des_cblock
7492 (which means, in function argument declarations, pointer to char)
7493 to des_cblock * (meaning pointer to array with 8 char elements),
7494 which allows the compiler to do more typechecking; it was like
7495 that back in SSLeay, but with lots of ugly casts.
7497 Introduce new type const_des_cblock.
7500 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
7501 problems: find RecipientInfo structure that matches recipient certificate
7502 and initialise the ASN1 structures properly based on passed cipher.
7505 *) Belatedly make the BN tests actually check the results.
7508 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
7509 to and from BNs: it was completely broken. New compilation option
7510 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
7511 key elements as negative integers.
7514 *) Reorganize and speed up MD5.
7515 [Andy Polyakov <appro@fy.chalmers.se>]
7518 [Richard Levitte <richard@levitte.org>]
7520 *) New option -out to asn1parse to allow the parsed structure to be
7521 output to a file. This is most useful when combined with the -strparse
7522 option to examine the output of things like OCTET STRINGS.
7525 *) Make SSL library a little more fool-proof by not requiring any longer
7526 that SSL_set_{accept,connect}_state be called before
7527 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
7528 in many applications because usually everything *appeared* to work as
7529 intended anyway -- now it really works as intended).
7532 *) Move openssl.cnf out of lib/.
7535 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
7536 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
7537 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
7538 [Ralf S. Engelschall]
7540 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
7541 handle PKCS#7 enveloped data properly.
7542 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
7544 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
7545 copying pointers. The cert_st handling is changed by this in
7546 various ways (and thus what used to be known as ctx->default_cert
7547 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
7548 any longer when s->cert does not give us what we need).
7549 ssl_cert_instantiate becomes obsolete by this change.
7550 As soon as we've got the new code right (possibly it already is?),
7551 we have solved a couple of bugs of the earlier code where s->cert
7552 was used as if it could not have been shared with other SSL structures.
7554 Note that using the SSL API in certain dirty ways now will result
7555 in different behaviour than observed with earlier library versions:
7556 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
7557 does not influence s as it used to.
7559 In order to clean up things more thoroughly, inside SSL_SESSION
7560 we don't use CERT any longer, but a new structure SESS_CERT
7561 that holds per-session data (if available); currently, this is
7562 the peer's certificate chain and, for clients, the server's certificate
7563 and temporary key. CERT holds only those values that can have
7564 meaningful defaults in an SSL_CTX.
7567 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
7568 from the internal representation. Various PKCS#7 fixes: remove some
7569 evil casts and set the enc_dig_alg field properly based on the signing
7573 *) Allow PKCS#12 password to be set from the command line or the
7574 environment. Let 'ca' get its config file name from the environment
7575 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
7579 *) Allow certificate policies extension to use an IA5STRING for the
7580 organization field. This is contrary to the PKIX definition but
7581 VeriSign uses it and IE5 only recognises this form. Document 'x509'
7585 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
7586 without disallowing inline assembler and the like for non-pedantic builds.
7589 *) Support Borland C++ builder.
7590 [Janez Jere <jj@void.si>, modified by Ulf Möller]
7595 *) SHA-1 cleanups and performance enhancements.
7596 [Andy Polyakov <appro@fy.chalmers.se>]
7598 *) Sparc v8plus assembler for the bignum library.
7599 [Andy Polyakov <appro@fy.chalmers.se>]
7601 *) Accept any -xxx and +xxx compiler options in Configure.
7604 *) Update HPUX configuration.
7607 *) Add missing sk_<type>_unshift() function to safestack.h
7608 [Ralf S. Engelschall]
7610 *) New function SSL_CTX_use_certificate_chain_file that sets the
7611 "extra_cert"s in addition to the certificate. (This makes sense
7612 only for "PEM" format files, as chains as a whole are not
7616 *) Support verify_depth from the SSL API.
7617 x509_vfy.c had what can be considered an off-by-one-error:
7618 Its depth (which was not part of the external interface)
7619 was actually counting the number of certificates in a chain;
7620 now it really counts the depth.
7623 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
7624 instead of X509err, which often resulted in confusing error
7625 messages since the error codes are not globally unique
7626 (e.g. an alleged error in ssl3_accept when a certificate
7627 didn't match the private key).
7629 *) New function SSL_CTX_set_session_id_context that allows to set a default
7630 value (so that you don't need SSL_set_session_id_context for each
7631 connection using the SSL_CTX).
7634 *) OAEP decoding bug fix.
7637 *) Support INSTALL_PREFIX for package builders, as proposed by
7641 *) New Configure options "threads" and "no-threads". For systems
7642 where the proper compiler options are known (currently Solaris
7643 and Linux), "threads" is the default.
7646 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
7649 *) Install various scripts to $(OPENSSLDIR)/misc, not to
7650 $(INSTALLTOP)/bin -- they shouldn't clutter directories
7651 such as /usr/local/bin.
7654 *) "make linux-shared" to build shared libraries.
7655 [Niels Poppe <niels@netbox.org>]
7657 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
7660 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
7661 extension adding in x509 utility.
7664 *) Remove NOPROTO sections and error code comments.
7667 *) Partial rewrite of the DEF file generator to now parse the ANSI
7671 *) New Configure options --prefix=DIR and --openssldir=DIR.
7674 *) Complete rewrite of the error code script(s). It is all now handled
7675 by one script at the top level which handles error code gathering,
7676 header rewriting and C source file generation. It should be much better
7677 than the old method: it now uses a modified version of Ulf's parser to
7678 read the ANSI prototypes in all header files (thus the old K&R definitions
7679 aren't needed for error creation any more) and do a better job of
7680 translating function codes into names. The old 'ASN1 error code imbedded
7681 in a comment' is no longer necessary and it doesn't use .err files which
7682 have now been deleted. Also the error code call doesn't have to appear all
7683 on one line (which resulted in some large lines...).
7686 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
7689 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
7690 0 (which usually indicates a closed connection), but continue reading.
7693 *) Fix some race conditions.
7696 *) Add support for CRL distribution points extension. Add Certificate
7697 Policies and CRL distribution points documentation.
7700 *) Move the autogenerated header file parts to crypto/opensslconf.h.
7703 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
7704 8 of keying material. Merlin has also confirmed interop with this fix
7705 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
7706 [Merlin Hughes <merlin@baltimore.ie>]
7708 *) Fix lots of warnings.
7709 [Richard Levitte <levitte@stacken.kth.se>]
7711 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
7712 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
7713 [Richard Levitte <levitte@stacken.kth.se>]
7715 *) Fix problems with sizeof(long) == 8.
7716 [Andy Polyakov <appro@fy.chalmers.se>]
7718 *) Change functions to ANSI C.
7721 *) Fix typos in error codes.
7722 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
7724 *) Remove defunct assembler files from Configure.
7727 *) SPARC v8 assembler BIGNUM implementation.
7728 [Andy Polyakov <appro@fy.chalmers.se>]
7730 *) Support for Certificate Policies extension: both print and set.
7731 Various additions to support the r2i method this uses.
7734 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
7735 return a const string when you are expecting an allocated buffer.
7738 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
7739 types DirectoryString and DisplayText.
7742 *) Add code to allow r2i extensions to access the configuration database,
7743 add an LHASH database driver and add several ctx helper functions.
7746 *) Fix an evil bug in bn_expand2() which caused various BN functions to
7747 fail when they extended the size of a BIGNUM.
7750 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
7751 support typesafe stack.
7754 *) Fix typo in SSL_[gs]et_options().
7755 [Nils Frostberg <nils@medcom.se>]
7757 *) Delete various functions and files that belonged to the (now obsolete)
7758 old X509V3 handling code.
7761 *) New Configure option "rsaref".
7764 *) Don't auto-generate pem.h.
7767 *) Introduce type-safe ASN.1 SETs.
7770 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
7771 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
7773 *) Introduce type-safe STACKs. This will almost certainly break lots of code
7774 that links with OpenSSL (well at least cause lots of warnings), but fear
7775 not: the conversion is trivial, and it eliminates loads of evil casts. A
7776 few STACKed things have been converted already. Feel free to convert more.
7777 In the fullness of time, I'll do away with the STACK type altogether.
7780 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
7781 specified in <certfile> by updating the entry in the index.txt file.
7782 This way one no longer has to edit the index.txt file manually for
7783 revoking a certificate. The -revoke option does the gory details now.
7784 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
7786 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
7787 `-text' option at all and this way the `-noout -text' combination was
7788 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
7789 [Ralf S. Engelschall]
7791 *) Make sure a corresponding plain text error message exists for the
7792 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
7793 verify callback function determined that a certificate was revoked.
7794 [Ralf S. Engelschall]
7796 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
7797 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
7798 all available cipers including rc5, which was forgotten until now.
7799 In order to let the testing shell script know which algorithms
7800 are available, a new (up to now undocumented) command
7801 "openssl list-cipher-commands" is used.
7804 *) Bugfix: s_client occasionally would sleep in select() when
7805 it should have checked SSL_pending() first.
7808 *) New functions DSA_do_sign and DSA_do_verify to provide access to
7809 the raw DSA values prior to ASN.1 encoding.
7812 *) Tweaks to Configure
7813 [Niels Poppe <niels@netbox.org>]
7815 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
7819 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
7822 *) New config option to avoid instructions that are illegal on the 80386.
7823 The default code is faster, but requires at least a 486.
7826 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
7827 SSL2_SERVER_VERSION (not used at all) macros, which are now the
7828 same as SSL2_VERSION anyway.
7831 *) New "-showcerts" option for s_client.
7834 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
7835 application. Various cleanups and fixes.
7838 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
7839 modify error routines to work internally. Add error codes and PBE init
7840 to library startup routines.
7843 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
7844 packing functions to asn1 and evp. Changed function names and error
7845 codes along the way.
7848 *) PKCS12 integration: and so it begins... First of several patches to
7849 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
7850 objects to objects.h
7853 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
7854 and display support for Thawte strong extranet extension.
7857 *) Add LinuxPPC support.
7858 [Jeff Dubrule <igor@pobox.org>]
7860 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
7861 bn_div_words in alpha.s.
7862 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
7864 *) Make sure the RSA OAEP test is skipped under -DRSAref because
7865 OAEP isn't supported when OpenSSL is built with RSAref.
7866 [Ulf Moeller <ulf@fitug.de>]
7868 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
7869 so they no longer are missing under -DNOPROTO.
7870 [Soren S. Jorvang <soren@t.dk>]
7873 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
7875 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
7876 doesn't work when the session is reused. Coming soon!
7879 *) Fix a security hole, that allows sessions to be reused in the wrong
7880 context thus bypassing client cert protection! All software that uses
7881 client certs and session caches in multiple contexts NEEDS PATCHING to
7882 allow session reuse! A fuller solution is in the works.
7883 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
7885 *) Some more source tree cleanups (removed obsolete files
7886 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
7887 permission on "config" script to be executable) and a fix for the INSTALL
7889 [Ulf Moeller <ulf@fitug.de>]
7891 *) Remove some legacy and erroneous uses of malloc, free instead of
7893 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
7895 *) Make rsa_oaep_test return non-zero on error.
7896 [Ulf Moeller <ulf@fitug.de>]
7898 *) Add support for native Solaris shared libraries. Configure
7899 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
7900 if someone would make that last step automatic.
7901 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
7903 *) ctx_size was not built with the right compiler during "make links". Fixed.
7906 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
7907 except NULL ciphers". This means the default cipher list will no longer
7908 enable NULL ciphers. They need to be specifically enabled e.g. with
7909 the string "DEFAULT:eNULL".
7912 *) Fix to RSA private encryption routines: if p < q then it would
7913 occasionally produce an invalid result. This will only happen with
7914 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
7917 *) Be less restrictive and allow also `perl util/perlpath.pl
7918 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
7919 because this way one can also use an interpreter named `perl5' (which is
7920 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
7921 installed as `perl').
7922 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7924 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
7925 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
7927 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
7928 advapi32.lib to Win32 build and change the pem test comparision
7929 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
7930 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
7931 and crypto/des/ede_cbcm_enc.c.
7934 *) DES quad checksum was broken on big-endian architectures. Fixed.
7937 *) Comment out two functions in bio.h that aren't implemented. Fix up the
7938 Win32 test batch file so it (might) work again. The Win32 test batch file
7939 is horrible: I feel ill....
7942 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
7943 in e_os.h. Audit of header files to check ANSI and non ANSI
7944 sections: 10 functions were absent from non ANSI section and not exported
7945 from Windows DLLs. Fixed up libeay.num for new functions.
7948 *) Make `openssl version' output lines consistent.
7949 [Ralf S. Engelschall]
7951 *) Fix Win32 symbol export lists for BIO functions: Added
7952 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
7953 to ms/libeay{16,32}.def.
7954 [Ralf S. Engelschall]
7956 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
7957 fine under Unix and passes some trivial tests I've now added. But the
7958 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
7959 added to make sure no one expects that this stuff really works in the
7960 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
7961 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
7963 [Ralf S. Engelschall]
7965 *) Fix the generation of two part addresses in perl.
7966 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
7968 *) Add config entry for Linux on MIPS.
7969 [John Tobey <jtobey@channel1.com>]
7971 *) Make links whenever Configure is run, unless we are on Windoze.
7974 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
7975 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
7979 *) Add a useful kludge to allow package maintainers to specify compiler and
7980 other platforms details on the command line without having to patch the
7981 Configure script everytime: One now can use ``perl Configure
7982 <id>:<details>'', i.e. platform ids are allowed to have details appended
7983 to them (seperated by colons). This is treated as there would be a static
7984 pre-configured entry in Configure's %table under key <id> with value
7985 <details> and ``perl Configure <id>'' is called. So, when you want to
7986 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
7987 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
7988 now, which overrides the FreeBSD-elf entry on-the-fly.
7989 [Ralf S. Engelschall]
7991 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
7994 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
7995 on the `perl Configure ...' command line. This way one can compile
7996 OpenSSL libraries with Position Independent Code (PIC) which is needed
7997 for linking it into DSOs.
7998 [Ralf S. Engelschall]
8000 *) Remarkably, export ciphers were totally broken and no-one had noticed!
8004 *) Cleaned up the LICENSE document: The official contact for any license
8005 questions now is the OpenSSL core team under openssl-core@openssl.org.
8006 And add a paragraph about the dual-license situation to make sure people
8007 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
8008 to the OpenSSL toolkit.
8009 [Ralf S. Engelschall]
8011 *) General source tree makefile cleanups: Made `making xxx in yyy...'
8012 display consistent in the source tree and replaced `/bin/rm' by `rm'.
8013 Additonally cleaned up the `make links' target: Remove unnecessary
8014 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
8015 to speed processing and no longer clutter the display with confusing
8016 stuff. Instead only the actually done links are displayed.
8017 [Ralf S. Engelschall]
8019 *) Permit null encryption ciphersuites, used for authentication only. It used
8020 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
8021 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
8025 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
8026 signed attributes when verifying signatures (this would break them),
8027 the detached data encoding was wrong and public keys obtained using
8028 X509_get_pubkey() weren't freed.
8031 *) Add text documentation for the BUFFER functions. Also added a work around
8032 to a Win95 console bug. This was triggered by the password read stuff: the
8033 last character typed gets carried over to the next fread(). If you were
8034 generating a new cert request using 'req' for example then the last
8035 character of the passphrase would be CR which would then enter the first
8039 *) Added the new `Includes OpenSSL Cryptography Software' button as
8040 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
8041 button and can be used by applications based on OpenSSL to show the
8042 relationship to the OpenSSL project.
8043 [Ralf S. Engelschall]
8045 *) Remove confusing variables in function signatures in files
8046 ssl/ssl_lib.c and ssl/ssl.h.
8047 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8049 *) Don't install bss_file.c under PREFIX/include/
8050 [Lennart Bong <lob@kulthea.stacken.kth.se>]
8052 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
8053 functions that return function pointers and has support for NT specific
8054 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
8055 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
8056 unsigned to signed types: this was killing the Win32 compile.
8059 *) Add new certificate file to stack functions,
8060 SSL_add_dir_cert_subjects_to_stack() and
8061 SSL_add_file_cert_subjects_to_stack(). These largely supplant
8062 SSL_load_client_CA_file(), and can be used to add multiple certs easily
8063 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
8064 This means that Apache-SSL and similar packages don't have to mess around
8065 to add as many CAs as they want to the preferred list.
8068 *) Experiment with doxygen documentation. Currently only partially applied to
8070 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
8071 openssl.doxy as the configuration file.
8074 *) Get rid of remaining C++-style comments which strict C compilers hate.
8075 [Ralf S. Engelschall, pointed out by Carlos Amengual]
8077 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
8078 compiled in by default: it has problems with large keys.
8081 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
8082 DH private keys and/or callback functions which directly correspond to
8083 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
8084 is needed for applications which have to configure certificates on a
8085 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
8087 For the RSA certificate situation is makes no difference, but
8088 for the DSA certificate situation this fixes the "no shared cipher"
8089 problem where the OpenSSL cipher selection procedure failed because the
8090 temporary keys were not overtaken from the context and the API provided
8091 no way to reconfigure them.
8092 The new functions now let applications reconfigure the stuff and they
8093 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
8094 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
8095 non-public-API function ssl_cert_instantiate() is used as a helper
8096 function and also to reduce code redundancy inside ssl_rsa.c.
8097 [Ralf S. Engelschall]
8099 *) Move s_server -dcert and -dkey options out of the undocumented feature
8100 area because they are useful for the DSA situation and should be
8101 recognized by the users.
8102 [Ralf S. Engelschall]
8104 *) Fix the cipher decision scheme for export ciphers: the export bits are
8105 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
8106 SSL_EXP_MASK. So, the original variable has to be used instead of the
8107 already masked variable.
8108 [Richard Levitte <levitte@stacken.kth.se>]
8110 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
8111 [Richard Levitte <levitte@stacken.kth.se>]
8113 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
8114 from `int' to `unsigned int' because it's a length and initialized by
8115 EVP_DigestFinal() which expects an `unsigned int *'.
8116 [Richard Levitte <levitte@stacken.kth.se>]
8118 *) Don't hard-code path to Perl interpreter on shebang line of Configure
8119 script. Instead use the usual Shell->Perl transition trick.
8120 [Ralf S. Engelschall]
8122 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
8123 (in addition to RSA certificates) to match the behaviour of `openssl dsa
8124 -noout -modulus' as it's already the case for `openssl rsa -noout
8125 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
8126 currently the public key is printed (a decision which was already done by
8127 `openssl dsa -modulus' in the past) which serves a similar purpose.
8128 Additionally the NO_RSA no longer completely removes the whole -modulus
8129 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
8131 [Ralf S. Engelschall]
8133 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
8134 BIO. See the source (crypto/evp/bio_ok.c) for more info.
8135 [Arne Ansper <arne@ats.cyber.ee>]
8137 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
8138 to be added. Now both 'req' and 'ca' can use new objects defined in the
8142 *) Add cool BIO that does syslog (or event log on NT).
8143 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
8145 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
8146 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
8147 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
8148 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
8151 *) Add preliminary config info for new extension code.
8154 *) Make RSA_NO_PADDING really use no padding.
8155 [Ulf Moeller <ulf@fitug.de>]
8157 *) Generate errors when private/public key check is done.
8160 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
8161 for some CRL extensions and new objects added.
8164 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
8165 key usage extension and fuller support for authority key id.
8168 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
8169 padding method for RSA, which is recommended for new applications in PKCS
8170 #1 v2.0 (RFC 2437, October 1998).
8171 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
8172 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
8173 against Bleichbacher's attack on RSA.
8174 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
8177 *) Updates to the new SSL compression code
8178 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8180 *) Fix so that the version number in the master secret, when passed
8181 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
8182 (because the server will not accept higher), that the version number
8183 is 0x03,0x01, not 0x03,0x00
8184 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8186 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
8187 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
8188 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
8191 *) Support for RAW extensions where an arbitrary extension can be
8192 created by including its DER encoding. See apps/openssl.cnf for
8196 *) Make sure latest Perl versions don't interpret some generated C array
8197 code as Perl array code in the crypto/err/err_genc.pl script.
8198 [Lars Weber <3weber@informatik.uni-hamburg.de>]
8200 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
8201 not many people have the assembler. Various Win32 compilation fixes and
8202 update to the INSTALL.W32 file with (hopefully) more accurate Win32
8206 *) Modify configure script 'Configure' to automatically create crypto/date.h
8207 file under Win32 and also build pem.h from pem.org. New script
8208 util/mkfiles.pl to create the MINFO file on environments that can't do a
8209 'make files': perl util/mkfiles.pl >MINFO should work.
8212 *) Major rework of DES function declarations, in the pursuit of correctness
8213 and purity. As a result, many evil casts evaporated, and some weirdness,
8214 too. You may find this causes warnings in your code. Zapping your evil
8215 casts will probably fix them. Mostly.
8218 *) Fix for a typo in asn1.h. Bug fix to object creation script
8219 obj_dat.pl. It considered a zero in an object definition to mean
8220 "end of object": none of the objects in objects.h have any zeros
8221 so it wasn't spotted.
8222 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
8224 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
8225 Masking (CBCM). In the absence of test vectors, the best I have been able
8226 to do is check that the decrypt undoes the encrypt, so far. Send me test
8227 vectors if you have them.
8230 *) Correct calculation of key length for export ciphers (too much space was
8231 allocated for null ciphers). This has not been tested!
8234 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
8235 message is now correct (it understands "crypto" and "ssl" on its
8236 command line). There is also now an "update" option. This will update
8237 the util/ssleay.num and util/libeay.num files with any new functions.
8239 perl util/mkdef.pl crypto ssl update
8240 it will update them.
8243 *) Overhauled the Perl interface (perl/*):
8244 - ported BN stuff to OpenSSL's different BN library
8245 - made the perl/ source tree CVS-aware
8246 - renamed the package from SSLeay to OpenSSL (the files still contain
8247 their history because I've copied them in the repository)
8248 - removed obsolete files (the test scripts will be replaced
8249 by better Test::Harness variants in the future)
8250 [Ralf S. Engelschall]
8252 *) First cut for a very conservative source tree cleanup:
8253 1. merge various obsolete readme texts into doc/ssleay.txt
8254 where we collect the old documents and readme texts.
8255 2. remove the first part of files where I'm already sure that we no
8256 longer need them because of three reasons: either they are just temporary
8257 files which were left by Eric or they are preserved original files where
8258 I've verified that the diff is also available in the CVS via "cvs diff
8259 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
8260 the crypto/md/ stuff).
8261 [Ralf S. Engelschall]
8263 *) More extension code. Incomplete support for subject and issuer alt
8264 name, issuer and authority key id. Change the i2v function parameters
8265 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
8266 what that's for :-) Fix to ASN1 macro which messed up
8267 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
8270 *) Preliminary support for ENUMERATED type. This is largely copied from the
8274 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
8275 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8277 *) Make sure `make rehash' target really finds the `openssl' program.
8278 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
8280 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
8281 like to hear about it if this slows down other processors.
8284 *) Add CygWin32 platform information to Configure script.
8285 [Alan Batie <batie@aahz.jf.intel.com>]
8287 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
8288 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
8290 *) New program nseq to manipulate netscape certificate sequences
8293 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
8297 *) Fixes to BN code. Previously the default was to define BN_RECURSION
8298 but the BN code had some problems that would cause failures when
8299 doing certificate verification and some other functions.
8300 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
8302 *) Add ASN1 and PEM code to support netscape certificate sequences.
8305 *) Add ASN1 and PEM code to support netscape certificate sequences.
8308 *) Add several PKIX and private extended key usage OIDs.
8311 *) Modify the 'ca' program to handle the new extension code. Modify
8312 openssl.cnf for new extension format, add comments.
8315 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
8316 and add a sample to openssl.cnf so req -x509 now adds appropriate
8320 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
8321 error code, add initial support to X509_print() and x509 application.
8324 *) Takes a deep breath and start addding X509 V3 extension support code. Add
8325 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
8326 stuff is currently isolated and isn't even compiled yet.
8329 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
8330 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
8331 Removed the versions check from X509 routines when loading extensions:
8332 this allows certain broken certificates that don't set the version
8333 properly to be processed.
8336 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
8337 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
8338 can still be regenerated with "make depend".
8341 *) Spelling mistake in C version of CAST-128.
8342 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
8344 *) Changes to the error generation code. The perl script err-code.pl
8345 now reads in the old error codes and retains the old numbers, only
8346 adding new ones if necessary. It also only changes the .err files if new
8347 codes are added. The makefiles have been modified to only insert errors
8348 when needed (to avoid needlessly modifying header files). This is done
8349 by only inserting errors if the .err file is newer than the auto generated
8350 C file. To rebuild all the error codes from scratch (the old behaviour)
8351 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
8352 or delete all the .err files.
8355 *) CAST-128 was incorrectly implemented for short keys. The C version has
8356 been fixed, but is untested. The assembler versions are also fixed, but
8357 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
8358 to regenerate it if needed.
8359 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
8360 Hagino <itojun@kame.net>]
8362 *) File was opened incorrectly in randfile.c.
8363 [Ulf Möller <ulf@fitug.de>]
8365 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
8366 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
8367 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
8368 al: it's just almost always a UTCTime. Note this patch adds new error
8369 codes so do a "make errors" if there are problems.
8372 *) Correct Linux 1 recognition in config.
8373 [Ulf Möller <ulf@fitug.de>]
8375 *) Remove pointless MD5 hash when using DSA keys in ca.
8376 [Anonymous <nobody@replay.com>]
8378 *) Generate an error if given an empty string as a cert directory. Also
8379 generate an error if handed NULL (previously returned 0 to indicate an
8380 error, but didn't set one).
8381 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
8383 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
8386 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
8387 parameters. This was causing a warning which killed off the Win32 compile.
8390 *) Remove C++ style comments from crypto/bn/bn_local.h.
8391 [Neil Costigan <neil.costigan@celocom.com>]
8393 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
8394 based on a text string, looking up short and long names and finally
8395 "dot" format. The "dot" format stuff didn't work. Added new function
8396 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
8397 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
8398 OID is not part of the table.
8401 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
8402 X509_LOOKUP_by_alias().
8405 *) Sort openssl functions by name.
8408 *) Get the gendsa program working (hopefully) and add it to app list. Remove
8409 encryption from sample DSA keys (in case anyone is interested the password
8413 *) Make _all_ *_free functions accept a NULL pointer.
8414 [Frans Heymans <fheymans@isaserver.be>]
8416 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
8418 [Anonymous <nobody@replay.com>]
8420 *) s_server should send the CAfile as acceptable CAs, not its own cert.
8421 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8423 *) Don't blow it for numeric -newkey arguments to apps/req.
8424 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
8426 *) Temp key "for export" tests were wrong in s3_srvr.c.
8427 [Anonymous <nobody@replay.com>]
8429 *) Add prototype for temp key callback functions
8430 SSL_CTX_set_tmp_{rsa,dh}_callback().
8433 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
8434 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
8437 *) X509_name_add_entry() freed the wrong thing after an error.
8438 [Arne Ansper <arne@ats.cyber.ee>]
8440 *) rsa_eay.c would attempt to free a NULL context.
8441 [Arne Ansper <arne@ats.cyber.ee>]
8443 *) BIO_s_socket() had a broken should_retry() on Windoze.
8444 [Arne Ansper <arne@ats.cyber.ee>]
8446 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
8447 [Arne Ansper <arne@ats.cyber.ee>]
8449 *) Make sure the already existing X509_STORE->depth variable is initialized
8450 in X509_STORE_new(), but document the fact that this variable is still
8451 unused in the certificate verification process.
8452 [Ralf S. Engelschall]
8454 *) Fix the various library and apps files to free up pkeys obtained from
8455 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
8458 *) Fix reference counting in X509_PUBKEY_get(). This makes
8459 demos/maurice/example2.c work, amongst others, probably.
8460 [Steve Henson and Ben Laurie]
8462 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
8463 `openssl' and second, the shortcut symlinks for the `openssl <command>'
8464 are no longer created. This way we have a single and consistent command
8465 line interface `openssl <command>', similar to `cvs <command>'.
8466 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
8468 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
8469 BIT STRING wrapper always have zero unused bits.
8472 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
8475 *) Make the top-level INSTALL documentation easier to understand.
8478 *) Makefiles updated to exit if an error occurs in a sub-directory
8479 make (including if user presses ^C) [Paul Sutton]
8481 *) Make Montgomery context stuff explicit in RSA data structure.
8484 *) Fix build order of pem and err to allow for generated pem.h.
8487 *) Fix renumbering bug in X509_NAME_delete_entry().
8490 *) Enhanced the err-ins.pl script so it makes the error library number
8491 global and can add a library name. This is needed for external ASN1 and
8492 other error libraries.
8495 *) Fixed sk_insert which never worked properly.
8498 *) Fix ASN1 macros so they can handle indefinite length construted
8499 EXPLICIT tags. Some non standard certificates use these: they can now
8503 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
8504 into a single doc/ssleay.txt bundle. This way the information is still
8505 preserved but no longer messes up this directory. Now it's new room for
8506 the new set of documenation files.
8507 [Ralf S. Engelschall]
8509 *) SETs were incorrectly DER encoded. This was a major pain, because they
8510 shared code with SEQUENCEs, which aren't coded the same. This means that
8511 almost everything to do with SETs or SEQUENCEs has either changed name or
8512 number of arguments.
8513 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
8515 *) Fix test data to work with the above.
8518 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
8519 was already fixed by Eric for 0.9.1 it seems.
8520 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
8522 *) Autodetect FreeBSD3.
8525 *) Fix various bugs in Configure. This affects the following platforms:
8529 unixware-2.0-pentium
8533 *) Eliminate generated files from CVS. Reorder tests to regenerate files
8534 before they are needed.
8537 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
8541 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
8543 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
8544 changed SSLeay to OpenSSL in version strings.
8545 [Ralf S. Engelschall]
8547 *) Some fixups to the top-level documents.
8550 *) Fixed the nasty bug where rsaref.h was not found under compile-time
8551 because the symlink to include/ was missing.
8552 [Ralf S. Engelschall]
8554 *) Incorporated the popular no-RSA/DSA-only patches
8555 which allow to compile a RSA-free SSLeay.
8556 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
8558 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
8559 when "ssleay" is still not found.
8560 [Ralf S. Engelschall]
8562 *) Added more platforms to Configure: Cray T3E, HPUX 11,
8563 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
8565 *) Updated the README file.
8566 [Ralf S. Engelschall]
8568 *) Added various .cvsignore files in the CVS repository subdirs
8569 to make a "cvs update" really silent.
8570 [Ralf S. Engelschall]
8572 *) Recompiled the error-definition header files and added
8573 missing symbols to the Win32 linker tables.
8574 [Ralf S. Engelschall]
8576 *) Cleaned up the top-level documents;
8577 o new files: CHANGES and LICENSE
8578 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
8579 o merged COPYRIGHT into LICENSE
8580 o removed obsolete TODO file
8581 o renamed MICROSOFT to INSTALL.W32
8582 [Ralf S. Engelschall]
8584 *) Removed dummy files from the 0.9.1b source tree:
8585 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
8586 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
8587 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
8588 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
8589 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
8590 [Ralf S. Engelschall]
8592 *) Added various platform portability fixes.
8595 *) The Genesis of the OpenSSL rpject:
8596 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
8597 Young and Tim J. Hudson created while they were working for C2Net until
8599 [The OpenSSL Project]
8602 Changes between 0.9.0b and 0.9.1b [not released]
8604 *) Updated a few CA certificates under certs/
8607 *) Changed some BIGNUM api stuff.
8610 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
8611 DGUX x86, Linux Alpha, etc.
8614 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
8615 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
8619 *) Add -strparse option to asn1pars program which parses nested
8621 [Dr Stephen Henson <shenson@bigfoot.com>]
8623 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
8626 *) DSA fix for "ca" program.
8629 *) Added "-genkey" option to "dsaparam" program.
8632 *) Added RIPE MD160 (rmd160) message digest.
8635 *) Added -a (all) option to "ssleay version" command.
8638 *) Added PLATFORM define which is the id given to Configure.
8641 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
8644 *) Extended the ASN.1 parser routines.
8647 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
8650 *) Added a BN_CTX to the BN library.
8653 *) Fixed the weak key values in DES library
8656 *) Changed API in EVP library for cipher aliases.
8659 *) Added support for RC2/64bit cipher.
8662 *) Converted the lhash library to the crypto/mem.c functions.
8665 *) Added more recognized ASN.1 object ids.
8668 *) Added more RSA padding checks for SSL/TLS.
8671 *) Added BIO proxy/filter functionality.
8674 *) Added extra_certs to SSL_CTX which can be used
8675 send extra CA certificates to the client in the CA cert chain sending
8676 process. It can be configured with SSL_CTX_add_extra_chain_cert().
8679 *) Now Fortezza is denied in the authentication phase because
8680 this is key exchange mechanism is not supported by SSLeay at all.
8683 *) Additional PKCS1 checks.
8686 *) Support the string "TLSv1" for all TLS v1 ciphers.
8689 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
8690 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
8693 *) Fixed a few memory leaks.
8696 *) Fixed various code and comment typos.
8699 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
8700 bytes sent in the client random.
8701 [Edward Bishop <ebishop@spyglass.com>]
projects / thirdparty / openssl.git / blob